21 CFR Part 11 Complete Guide To International Computer Validation Compliance PDF

21 CFR Part 11: Complete Guide to
International Computer Validation

Compliance for the Pharmaceutical
Industry
21 CFR Part 11: Complete Guide

to International Computer
Validation Compliance for the
Pharmaceutical Industry
Orlando Lpez
Boca Raton London New York Washington, D.C.
This edition published in the Taylor & Francis e-Library, 2005.

To purchase your own copy of this or any of Taylor & Francis or Routledge's collection of
thousands of eBooks please go to www.eBookstore.tandf.co.uk.
Library of Congress Cataloging-in-Publication Data
Lpez, Orlando. 21 CFR Part 11: complete guide to international computer validation compliance
for the pharmaceutical industry/Orlando Lpez. p. cm. Includes bibliographical references and
index. ISBN 0-8493-2243-X (alk. paper) 1. Pharmaceutical industry. 2. Computer software
Validation, 3. Computer programsVerification. I. Title: Twenty-one CFR Part Eleven. II. Title.
RS192.L67 2004 338.4'76151dc22 2003063460
This book contains information obtained from authentic and highly regarded sources. Reprinted
material is quoted with permission, and sources are indicated. A wide variety of references are
listed. Reasonable efforts have been made to publish reliable data and information, but the author
and the publisher cannot assume responsibility for the validity of all materials or for the
consequences of their use.
Neither this book nor any part may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, microfilming, and recording, or by any
information storage or retrieval system, without prior permission in writing from the publisher.
The consent of CRC Press LLC does not extend to copying for general distribution, for promotion,
for creating new works, or for resale. Specific permission must be obtained in writing from CRC
Press LLC for such copying.
Direct all inquiries to CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and
are used only for identification and explanation, without intent to infringe.
Visit the CRC Press Web site at www.crcpress.com
2004 by CRC Press LLC
No claim to original U.S. Government works
ISBN 0-203-02677-2 Master e-book ISBN
International Standard Book Number 0-8493-2243-X (Print Edition)

Library of Congress Card Number 2003063460
Contents
Foreword
by Sion Wyn
Preface
by Orlando Lpez
Publishers Note
vii
Dedication
xiii
ix
xi
Chapter 1 Introduction
Chapter 2 Validation Overview
Chapter 3 USA Regulatory Requirements for Computer Systems
15
Chapter 4 New Computer Systems Validation Model
24
Chapter 5 Computer Validation Management Cycle
28
Chapter 6 Computer Validation Program Organization
32
Chapter 7 The Computer Systems Validation Process
38
Chapter 8 Validation Project Plans and Schedules
46
Chapter 9 Inspections and Testing
53
Chapter 10 Qualifications
62
Chapter 11 SLC Documentation
90
Chapter 12 Relevant Procedural Controls
95
Chapter 13 Change Management
97
Chapter 14 Training
101
Chapter 15 Security
103
Chapter 16 Source Code
117
Chapter 17 Hardware/Software Suppliers Qualification
120
Chapter 18 Maintaining the State of Validation
124
Chapter 19 Part 11 Remediation Project
131
Chapter 20 Operational Checks
136
Chapter 21 Compliance Policy Guide (CPG) 7153.17
140
Chapter 22 Electronic Records
144
Chapter 23 Electronic Signatures
153
Chapter 24 Technologies Supporting Part 11
158
Chapter 25 All Together
166
Chapter 26 The Future
173
A Glossary of Terms
177
B Abbreviations and Acronyms
187
C Applicability of a Computer Validation Model
189
D Criticality and Complexity Assessment
197
E Sample Development Activities Grouped by Project Periods
206
F Administrative Procedures Mapped to Part 11
235
G Sample Audit Checklist for a Closed System
241
H Computer Systems Regulatory Requirements
244
I Technical Design Key Practices

Index
267
269
Foreword
It has been my pleasure to know Orlando Lpez as a friend for some years, and it has also
been my privilege to work with him in his specialty field, computer systems validation
and compliance. Orlando is well known in the industry as an enthusiastic advocate and
supporter of industry initiatives such as GAMP, and a valued contributor to various task
teams and committees.
He is a practitioner with valuable experience in tackling real challenges and dealing
with real-life problems. In this book, he illuminates the role of quality assurance, and
shows the importance of integrating the validation activities into the system life cycle
within a structured top-down approach.
This book reflects Orlandos breadth of knowledge of and experience in the regulated
pharmaceutical and associated healthcare industries, software engineering, and quality
assurance. It covers the regulatory requirements, and the organisation, planning,
verification, and documentation activities required to meet those requirements. It also
describes the administrative and procedural controls required for compliance, as well as
introducing appropriate supporting technologies such as encryption and digital signatures.
Orlando is enthusiastic about his subject, and he is a mine of useful up-to-the-minute
information. He is always ready and willing to share this information and experience with
others for the benefit of the industry and the patient.
This book shows his commitment to the practical application of quality assurance and
engineering techniques in the development of systems that meet user and regulatory
requirements.
Sion Wyn
Conformity Limited
United Kingdom
Preface
This book speaks to all those involved in planning and undertaking computer systems
validation in the United States of Americas Food and Drug Administration (FDA)
regulated industries, mainly pharmaceuticals. The topics and discussion levels reflect
what I believe such people will find interesting or useful. My vision for this book is that it
reveals my perspective on how to perform computer systems validation, and gives my
readers a clear picture on how they must work effectively in a quality role in todays
high-pressure, regulated environment.
Focusing on project management, I have chosen to include materials not normally
stressed in a computer systems book applicable to the FDA regulated industries.
Estimating, planning, and scheduling computer systems validation are hard to pin down,
because there is no true end to a project. More verification and testing are always
required, and there are serious risks associated with skipping them. Project managers
must be fully aware of these trade-offs, and need to share this knowledge with their
teams. Efficiency is a major concern in this book.
Computer systems validation personnel must also deal with design errors. A program
that perfectly meets a lousy specification is a lousy program. Specifically for medical
devices, books on software reliability tend to set aside the user interface issue, and treat it
as the sole province of the human factor analyst. The reliability of a system is determined
by how all its various parts, including the people who use it, work together.
Your role as a project manager is to find and highlight problems in a product, with the
aim of improving its quality. You are one of the few who will have the required
information, or be directly able to examine the product in close detail, before it is
transferred to the regulated environment sector.
In the FDA-regulated environment, it is not acceptable for programmers or their
managers to deviate from standardized methodologies. Control programs for
manufacturing facilities must be well documented and thoroughly specified. Computer
systems must only change after careful control, since the electronic records associated
within these systems are now highly relevant to new FDA regulations. If manufacturing
systems fail, the quality of the product is in question: this can be catastrophic to the users,
patients whose lives depend largely on quality assurance staffs, and on huge computer
systems budgets being met.
It is fascinating how, in many highly standardized environments, a computer systems
validation (CSV) group works with the product just before it goes to production, or when
it has just been put into internal use.
This group is not involved in the drug development effort, their budgets are often tiny,
and their deadlines almost impossible to achieve. The quality assurance organization may
look down its collective nose at them and the CSV team is not involved with a product
for the time required to perform real product life-cycle verification and testing.
With the introduction of 21 CFR Part 11 by the FDA, CSV practitioners are
encountering a new wave of computer systems validation. The computer validation
professional needs to be aware of quality engineering, project management, and computer

technologies. This book emphasizes the promotion of a conceptual understanding of
computer systems validation, and the need for evidence of such understanding in the form
of practical examples.
Commencing with a discussion of the regulatory requirements associated with the
practical foundation work for any project, this book introduces a model to determine the
Part 11 requirements for consideration during implementation; this is based on essential
characteristics of computer systems. This book then establishes a top-down approach to
the integration of computer systems validation strategies in support of a computer
systems validation project; focussing on the practical issues in computer systems
implementation and operational life project management; including relevant activities to
comply with Part 11, validation planning, and scheduling. Inspections and testing,
computer systems qualification, and sample qualification using the Good Automated
Manufacturing Practice (GAMP) categories of software, are all covered here.
The second section reviews documentation in relation to CSV and the procedural
controls required for regulated operations. Change management and control, training, and
security (introducing key technology-driven services such as user/data authentication and
access control), are followed by guidance on source code issues, and suppliers
qualifications.
The book explains how technologies such as hashing, encryption, and digital
signatures can support Part 11. These technologies have not currently been implemented
to any extent in the FDA-regulated environments.
Finally, my vision of the integration of project management and computer systems
validation is based on hands-on experience and the knowledge accumulated from more
than 20 years experience in software engineering and software quality engineering. The
author wishes to acknowledge and express appreciation to all of the people who, during
the last 30 years, have contributed to this body of knowledge, and who laid down the
crucial groundwork for this computer systems validation book.
Orlando Lpez, Executive Consultant
J&J Networking and Computing Services
Raritan, New Jersey
Publishers Note
The FDA withdrew its last guidance document in the series on electronic copies of
electronic records in February 2003. But it did NOT withdraw the 21 CFR Part 11 Rule,
which is still in force, and must be complied with: the FDA is still inspecting companies
for compliance and none of the multi-million dollar fines imposed by the FDA for noncompliance have been revoked. Nor have any of the preceding warning letters been
recalled.
This Professional Edition of a pragmatic guide is designed to enlighten the industry at
large, which suffers somewhat from the confusion created by earlier guidance documents
now withdrawn. Written by an international expert in computer software and computer
validation, it will become a classic of its kind when properly used to help companies
implement their compliance requirements to meet the FDA and the MCA.
On September 3, 2003 the FDA released the final guidance on the Scope and
Application of 21 CFR Part 11. Comments from industry played a large role in the
changes that the FDA made between the draft version and the new final version.
Significant clarification was also established on the definition of legacy systems and on
enforcement discretion.
Despite these clarifications, there is still a great deal of uncertainty remaining
regarding what is expected by the Agency in the long term and how to proceed
logistically with a risk-based remediation for Part 11. To access these guidelines, please
visit the following website: http:/www.fda.gov/cder/guidance/5667fnl.htm.
For Lizette, Mikhail, Istvn, and Christian
Chapter 1
Introduction
FDA regulations require that domestic or foreign manufacturers of regulated products
intended for commercial distribution in the United States, establish and follow controls as
part of a quality assurance (QA) program. These controls help to ensure that regulated
products are safe and effective. The QA controls and associated records required by the
FDA are contained in the applicable predicate regulations.
The records-keeping requirements to demonstrate the safety and efficacy of products
are also applicable for persons exporting human drugs, biological products, devices,
animal drugs, food, and cosmetics that may not be marketed or sold in the United States.1
Computer systems used to collect and manage records must demonstrate to the FDA
that regulated products are safe and efficacious. Increased attention being paid to
computers systems by the regulatory authorities requires that a periodic evaluation of
compliance must be carried out by computer systems developers,2 companies, industry,
and/or the regulator.
As with any production equipment performing regulated operations, computer systems
must be designed and deployed in compliance with specific performance and quality
standards. In the FDA regulation context, validation is the process of proving, in
accordance with the principles of the applicable predicate regulations, that any procedure,
process, equipment, materials, activity, or system actually leads to the expected result.
The documentation generated during the validation can be subject to examination by
FDA field investigators. The results of a highquality validation program can ensure a
high degree of assurance of the trustworthiness of the electronic records and computer
system functionality.
Computer systems validation, as established in 21 CFR Part 211.68, Automatic,
Mechanical, and Electronic Equipment, is one of the most important requirements in
FDA-regulated operations and an element of the system life cycle (SLC). In addition to
the testing of the computer technology, other verifications and inspection activities
include code walkthroughs, dynamic analysis and trace analysis. These activities may
require 40% of overall project efforts.
In common with all FDA-regulated products, quality is built into a computer system
during its conceptualization, development, and operational life. The validation of
computer systems is an ongoing process that is integrated into the entire SLC.
Federal Register: December 19, 2001 (Vol. 66, No. 244).

Note. In this book a developer can either be an external company or an in-house software
development group.
2
21 CFR Part 11
Given the type of software commonly found in FDA-regulated operations, the

criticality and complexity of computer systems, the evolving regulatory climate, and the
current industry best practices, no single fixed standard can be applied to computer
systems validation. However, 21 CFR Part 11; Electronic Records, Electronic Signatures
Rule (hereafter referred to as Part 11), effective since August 1997, provides the explicit
and current regulatory trends applicable to computer systems performing regulated
operations. Only electronic records and electronic signatures that meet Part 11 can be
used to satisfy the requirements in the applicable predicate rule.
A computer systems validation program for regulated operations can be established
based on the regulatory requirements, SQA practices, SQE practices, and type of
software.
This book, relevant to FDA-regulated operations, provides practical information to
enable compliance with computer systems validation requirements, while highlighting
and efficiently integrating the Part 11 requirements into the computer validation program.
The ideas presented in this book are based on many years of experience in the United
States Department of Defense and FDA-regulated industries in various computer systems
development, maintenance, and quality functions.
The practical approach we take is intended to increase efficiency and ensure that
correct software development and maintenance is achieved. The topics addressed in this
book are equally relevant to other type of software development industries. It is not the
intention of this book to develop a paradigm or model for the regulated industry.
Chapter 2 introduces the reader to the computer systems validation concepts in the
context of regulated operations.
Chapter 3 discusses the regulatory requirements and furnishes the reader with the
foundations of the CSV approach in this book.
Chapter 4 introduces a model to help determine the Part 11 requirements to be
considered during implementation. It focuses on the basic characteristics of computer
systems. Practical examples on how to implement this model are provided in Appendix
C.
Chapter 5 establishes the top-down approach for integrating computer systems
validation strategies in support to the computer systems validation program.
Chapter 6 suggests an organization structure supporting the computer validation
program.
Chapter 7 is the focus of this book, being a practical approach to computer systems
implementation and operational life project management, and including the activities
relevant for compliance with Part 11. Appendix E expands the concept presented in this
chapter.
Chapter 8 covers validation planning and scheduling.
Chapter 9 discusses the subject of inspections and testing as part of the computer
validation process.
Chapter 10 covers the subject of computer systems qualification, including sample
qualification using the GAMP software categories.
Chapter 11 discusses the SLC documentation and its relationship with computer
systems validation.
Chapter 12 covers the issues of procedural controls in regulated operations.
Change management and training are covered in Chapters 13 and 14, respectively.
Introduction
Chapter 15 discusses security, the backbone of Part 11. This chapter introduces the
key security services, user or data authentication, and access control.
Chapter 16 covers the issue of Source Code.
Chapter 17 covers supplier qualifications.
Chapter 18 shows how to maintain the state of the validation in computer systems.
After the system has been released for operation, the maintenance activities take over.
Chapter 19 briefly discusses Part 11 remediation activities.
Chapter 20 covers the important subject of Operational Systems Checks.
Chapter 21 covers the most recent Compliance Policy Guide 7153.17 applicable to
computer systems. The scope of this guide is Part 11.
Chapter 22 addresses specific issues on electronic records.
Chapter 23 covers the subject of electronic signatures and their implementation based
on the requirements of Part 11.
Chapter 24 discusses how hashing, encryption, and digital signature technologies can
be used to support Part 11.
Chapter 25 guides the reader on how to put together all the elements discussed in this
book.
Chapter 26 introduces the reader to the future of software engineering and suggests
how QA and QE must adapt to these changes.
Chapter 2
Validation Overview
WHAT IS A COMPUTER SYSTEM?
The term computer system can define any of the following: desktop systems; client or
of the following: desktop systems; client or server systems; automated process control
and laboratory systems; host-based systems; data acquisition and analysis systems; and
all associated software. The associated software comprises application software or
firmware, system software, and computer system supporting documentation.
Regulatory guidance
Computer systems shall be validated. The computer validation must ensure
accuracy, reliability, consistent intended performance, and the ability to discern
invalid or altered records.
Department of Health and Human Services, Food and Drug Administration, 21
CFR Part 11, Electronic Records; Electronic Signatures, Federation Register 62
(54), 1343013466, March 20, 1997.
In the regulatory context, computer systems are integrated into the operating environment
(Figure 211). The operating environment may include the process or operation being
controlled or monitored by the computer system, the procedural controls, process-related
documentation, and the people. Computer systems performing regulated operations may
either control the quality of a product during its development, testing, manufacturing, and
handling processes; manage information business operations; manage data used to prove
the safety; efficacy and quality of the product and formulation; and provide data for drug
submissions.
WHAT IS A COMPUTER SYSTEMS VALIDATION?

The validation of computer systems performing regulated operations provides
1
R.R.Herr and M.L.Wyrick, A Globally Harmonized Glossary of Terms for Communicating

Computer Validation Key Practices, PDA Journal of Pharmaceutical Science and Technology,
March/April 1999.
21 CFR Part 11
confirmation by examination and provision of objective evidence that computer system

specifications conform to user needs and intended uses, and that all requirements can be
consistently fulfilled.2
It encompasses:
Figure 21. A Computer System and

the Operating Environment.
planning, verification, testing, traceability, configuration management, and many other
aspects of good software engineeringthat together help to support a final conclusion
that software is validated.3
The key words in the preceding paragraph are objective evidence.
Objective evidence is generated following completion of a number of formal and
informal activities, many of which must be completed in a predefined order. These
activities comprise the SLC. The work products of each phase in the SLC provide the
objective evidence that is required to demonstrate that the computer system conforms to
the needs and intended uses of the user, and that all requirements can be consistently
fulfilled.
2
FDA, draft of Guidance for the Industry: 21 CFR Part 11; Electronic Records; Electronic
Signatures: Glossary of Terms, August 2001.
3
FDA CDRH, General Principles of Software Validation; Final Guidance for Industry and FDA
Staff, January 2002.
Validation overview
WHY DO WE VALIDATE COMPUTER SYSTEMS?

Computer systems are validated for two important reasons: to ensure that good business
practices are followed, and to satisfy regulatory agency requirements.
The main business reasons to validate computer systems are to demonstrate
conformance with the system requirements specification, to increase acceptance of the
systems by end-users, and to avoid high maintenance costs. Regarding maintenance costs,
Figure 22 depicts the relative cost in 1976 of repairing software. At present, the
proportions are similar. The longer a defect is uncovered, the more expensive it is to
repair.
Figure 22. Relative Cost of

Repairing Software Error.4
Taking as an example the conventional waterfall development methodology, an error
introduced during the technical design activity will cost twice as much to fix during the
program build stage. Each error uncovered during the operational life period will cost 20
times more to fix it during the coding stage.
Up to 50% of maintenance effort is spent fixing errors that have been present since the
initial phase of the development.
The effect that verification and testing activities have on financial resources is
noticeable. A potential saving of 50% on software maintenance can make a sizeable
impact on overall life cycle cost.
4
B.W.Boehm, Software Engineering, IEEE Transactions on Computers, 1976.
21 CFR Part 11
Good business practices include assurance of the trustworthiness of the data and
information.
Other reasons to satisfy good business practices include:
demonstrating the suitability of computer hardware and software to perform the
assigned task
facilitating the relationships between the company and its suppliers
improving the level of control and maintainability
The reasons for satisfying regulatory agency requirements include:
minimizing regulatory actions
maintaining a positive relationship with regulatory agencies
expediting submissions to and approval by the FDA
avoiding product recalls and negative publicity
KEY PROJECT ELEMENTS

Key elements that must be successfully executed on computer implementation projects
are:
selection of a development, deployment, and/or maintenance methodology that best
suits the nature of the project
establishment of intended use and proper performance of the system
selection of hardware based on capacity and functionality
identification and testing of the operational limits to establish production procedures
identification of operational functions associated with the users, sequencing,
regulatory, security, company (e.g., standards), safety requirements, and so on
identification and testing of worst-case operational or production conditions
accuracy and reproducibility of the testing results
project documentation requirements5
written design specification6 that describes what the software is intended to do and
how it will do it
a written qualification/test plan based on the design specification,7 including both
structural and functional analysis, as applicable
test results and an evaluation of how these results demonstrate that the
predetermined design specification has been met (e.g., requirements traceability
analysis)
5
FDA, Guidance for the Industry: Computerized Systems Used in Clinical Trials, April 1999.
In this context a design specification may include a set of specifications containing the project
requirements, technical, configuration, functional, and/or technical design.
7
FDA, Guidance for the Industry: Computerized Systems Used in Clinical Trials, April 1999.
6
Validation overview
availability of procedural controls to maintain the operating environment of the system

environments
any modification to a component of the system and its operating environment must be
evaluated to determine the impact on the system. If necessary qualification or
validation will have to be re-executed, partially or totally
WHICH SYSTEMS SHOULD BE VALIDATED?

Computer systems performing regulated operations shall be validated. Typical examples
of such systems are:
systems that control the quality of product during its development, manufacturing,
testing, and handling processes
systems that create, modify, maintain, archive, retrieve, or transmit data used to prove
the safety, efficacy, and quality of product and formulations
systems that provide information on which decisions are made affecting fitness for
purpose of the product
systems that create, modify, maintain, archive, retrieve, or transmit records that must be
available for inspection by a regulatory agency
systems that provide data or reports that are sent to regulatory agencies
The depth and scope of the validation depends on the category of the software, and the
complexity and criticality of the application. Refer to Chapter 7 for details of the
computer system validation process.
INTRODUCTION TO THE COMPUTER SYSTEMS VALIDATION

PROCESS
The system life cycle (SLC) is the period of time that begins when a product is
conceived and ends when the product is no longer available for use.8
The development model associated with the SLC contains the software engineering
tasks and associated work products necessary to support the computer system validation
effort. It breaks the systems development process down into sub-periods during which
discrete work products are developed. This approach leads to well-documented systems
that are easier to test and maintain, and for which an organization can have confidence
that the systems functions will be fulfilled with a minimum of unforeseen problems.
The development model contains specific inspection and testing tasks that are
appropriate for the intended use of the computer system.
The naming of each sub-period depends on the development model used. The
following enumerates development methodologies common in the FDA-regulated
industry.
8
ANSI/IEEE Std 610.121990, Standard Glossary of Software Engineering Terminology, Institute
of Electrical and Electronic Engineers, New York, 1990.
21 CFR Part 11
10
Classical Life CycleSystem Engineering, Analysis, Design, Code, Testing and

Maintenance
PDA/PhRMAs Computer System Validation CommitteeSystem Definition, Design,
Unit Testing, Integration Computerized System, System Module Integration, System
(Total) Integration, Operation, and Maintenance
IEEEConcept, Requirements, Design, Implementation, Test, Installation and
Checkout, Operation and Maintenance
FDA9Requirements, Design, Implementation, Test, Installation and Checkout,
Operation and Maintenance
CDRH10Specification/Requirements, Design, Implementation, Verification and
Validation, Maintenance
FDA/NCTR11Initiation, Requirements Analysis, Design, Programming and Testing,
System Integration and Testing, System Validation, System Release, Operation and
Maintenance
EU Annex 11Planning, Specification, Programming, Testing, Commissioning,
Documentation, Operation, Monitoring and Modifying
DOD-STD-2167ASystem Requirements Analysis/Design, Software Requirements
Analysis, Preliminary Design, Detailed Design, Coding and Computer System Unit
Testing, Computer Software Component Integration and Testing, Computer Software
Configuration Item Testing, System Integration and Testing
9
FDA, Software Development Activities, Reference Materials and Training Aids for Investigation,
July 1987.
10
FDA, Review Guidance for Computer Controlled Medical Devices Undergoing 510(k) Review,
Office of Device Evaluation Center for Devices and Radiological Health, 1991.
11
Office of Regulatory Affairs of the Food and Drug Administration and the National Center for
Toxicological Research, Computerized Data Systems for Nonclinical Safety Assessment, October
11, 1987.
Validation overview
11
Figure 23. System Life Cycle

focusing on Software Engineering
Key Practices.
Fourth Generation TechniquesRequirements gathering, Design strategy,
Implementation using 4GL, Product
Object OrientedThe object-oriented (OO) model stresses the notion of software
component reuse and comprises: requirements gathering, analysis, design,
implementation (browse library or creation of reusable components), prototype build,
and testing. The creation of reusable components is a development model within the
OO model, and comprises the following: specification, design, code, unit test,
integration into the prototype, and integration testing
Figure 2312 depicts an SLC adapted to different system acquisition strategies and
software development models. It is focused on software engineering key practices. It does
not specify or discourage the use of any particular software development method. The
acquirer determines which of the activities outlined by the standard will be conducted,
and the developer is responsible for selecting methods that support the achievement of
contract requirements. A modifiable framework must be tailored to the unique
characteristics of each project. It includes the following periods:
conceptualization
development
12
R.R.Herr and M.L.Wyrick, A Globally Harmonized Glossary of Terms for Communicating

Computer Validation Key Practices, PDA Journal of Pharmaceutical Science and Technology,
March/April 1999.
21 CFR Part 11
12
early operational life

maturity
aging
Project Recommendation, Project Initiation, Release for Use, and Retirement are events
that must be considered as phase gates or major decision points, and that include formal
approvals before development can proceed to the next period.
Certain discrete work products are expected when evidencing the development and
maintenance work of computer systems compliance to regulatory requirements. The
selected SLC should specify the overall periods and the associated events, and dictate the
minimum requirements regardless of the chosen development method. These
development methods and deliverables are also known as the development methodology.
The development methodology establishes detailed discrete work products, by phases
and events, and by associated activities. The computer systems development approach
must be consistent with the selected SLC; the most common development methodologies
are the Waterfall Model, the Incremental Development Model, the Evolutionary Model,
the Object Oriented approach, and the Spiral Model. Sample SLC activities and work
products are detailed in Appendix E.
A critical function of the validation process is to provide an assurance that the
development and operational life methodologies are consistent and followed carefully.
The SLC and its associated development and operational life methodologies, applicable
to computer systems performing regulated operations, shall be described in procedural
control(s). The project team may have the authority to select a developmental and
operational life methodology best suited to the nature of the system under development,
and operational life methodologies that differ from those originally included in the related
procedural control. In this case, the selected development or operational life
methodologies must be clearly and satisfactorily explained in the validation plan or
equivalent document.
It is the objective of FDA regulated companies in general to select the appropriate
SLC and associated methodologies. Development and maintenance teams must receive
adequate training in the use of the chosen methodology
To evaluate adherence to the selected methodology, quality checkpoints (e.g., audits)
are conducted during the project. Considering Figure 23 as an example, each event
associated when the SLC is a checkpoint when quality checks can be conducted.
Detailed descriptions of the infrastructure for supporting computer systems validation
can be found in Chapters 5 and 6. A sample computer systems development and
maintenance methodology can be found in Chapter 7.
COMPUTER SYSTEMS VALIDATION FOR LOW-CRITICALITY

AND/OR LOW COMPLEXITY PROJECTS
The methods contained in this book apply to any type of computer systems validation
project. However, for low criticality and/or low complexity projects, certain steps in the
computer systems project can be abbreviated: documentation may be shortened and
combined; reviews can be less formal, but not less thorough. Appendix D, Sample
Validation overview
13
Criticality and Complexity Assessment, provides guidelines for the management of and
approach to low-criticality or low-complexity projects.
Chapter 3
USA Regulatory Requirements for
Computer Systems1
The FDAs authority to regulate computer system functions in a regulated operation is
derived from the FD&C Act.2 Resulting from the FD&C Act, three (3) main items were
influential in computer systems practices in the regulated operations: applicable predicate
rule, CPGs, and Part 11.
In the predicate rules arena, good manufacturing practices (GMPs) applicable to the
manufacture of pharmaceuticals and medical devices are the most crucial regulations to
computer system practices in the regulated operations. GMPs are the specific regulations
describing the methods, equipment, facilities, and controls required for producing human
and veterinary products, medical devices, and processed food. The regulatory authority
enforcing the GMPs is the FDA.
In the United States, drug GMPs were formally introduced in 1963 and included the
regulation of computer systems. 21 CFR Part 211.2(b) emphasized computer backups and
documentation, including keeping hardcopy of master formulas, specifications, test
records, master production and control records, batch production records (batch
production and control records), and calculations.
By 19783 the regulations combined 21 CFR Part 211.2(b) and 211.68. The product of
this combination was the updated 21 CFR Part 211.68 (Automatic, mechanical, and
electronic equipment).
In summary, 21 CFR Part 211.68 requires that:
There must be a written program detailing the maintenance of the computer system
There must be a system to control changes to the computer hardware and software,
including documentation
There must be documented checks of I/Os for accuracy. In practice, it is implied that all
computer systems under the regulations must be qualified/validated
1
Lpez, O., FDA Regulations of Computer Systems in Drug Manufacturing13 Years Later,
Pharmaceutical Engineering, Vol. 21, No. 3, May/June 2001.
2
Section 510(a) (2)(B) of the Federal Food, Drug, and Cosmetic Act.
3
FR, Vol. 43, No. 190, September 29, 1978.
21 CFR Part 11
16
There must be programs to ensure the accuracy and security of computer inputs,
outputs, and data
Computer electronic records must be controlled, and this includes record backup,
security, and retention
Computer systems validation, implied in 21 CFR Part 211.68, established in 21 CFR Part
11.10(a), and defined in the recent draft FDA guideline,4 is one of the most important
requirements applicable to computer systems performing regulated operations. Computer
systems validation is the confirmation (by examination and the provision of objective
evidence) that computer system specifications conform to user needs and intended uses,
and that all requirements can be consistently fulfilled. It involves establishing that the
computer system conforms to the user, regulatory, safety, and intended functional
requirements.
Computer systems validation is an element of the SLC. In addition to the software and
hardware testing, other verification activities include code walkthroughs, dynamic
analysis, and trace analysis.
To clarify the intent of 21 CFR Part 211.68, CPG5 7132a.07, I/O Checking was
published in 1982. According to this CPG, computers I/Os must be tested for data
accuracy as part of the computer system qualification and, after the qualification, as part
of the computer systems ongoing verification program.
CPG 7132a.07 is based on the realistic anticipation that computer I/O errors can occur
on validated systems. Computer components such as logic circuits, memory,
microprocessors, or devices such as modems and displays, can fail after they have been
tested, just like any other mechanical part. Other sources of computer systems I/O
malfunctions are electromagnetic interferences, e.g., radio-frequency interference,
electrostatic discharge, and power disturbance. Software errors undetected during the
validation process may also be a source of I/O errors. In order to detect errors before a
computer system makes decisions using tainted data, an on-going monitoring program
needs to be established and followed to verify hardware and software I/Os during the
operation of the system.
The level, frequency, and extent of the I/O checking was suggested in the Federal
Register of January 20, 1995 (60 FR 4091). The level and frequency of the I/O
verifications must be guided by written procedure, and based on the complexity and
reliability of the computer system.
Note: A CPG as described in 21 CFR 10.85 is considered an Advisory Opinion
directed to FDA inspectors. These guidelines are the mechanisms utilized by the FDA to
spread policy statements within the Agency and to the public.
Another topic relevant to computer systems in GMP environments is the management
of application source code. Before 1985, 80% of computer systems were custom-built,
making the source code a deliverable. Since then, software developers have begun to
make off-the-shelf applications available: today, 80% of applications utilized to supervise
4
FDA, draft of Guidance for Industry: 21 CFR Part 11; Electronic Records; Electronic Signatures
Glossary of Terms, August 2001.
5
Lpez, O., A Guide to Computer Compliance Guides, Journal of cGMP Compliance, January
1997.
USA regulatory requirements for computer systems
17
manufacturing processes are configurable software. Many of the custom-built

programs can be found, for example, in PLC. For configurable applications, the source
code is the proprietary information of the developer, so the likelihood of acquiring the
source code is very low.
The regulatory requirements applicable to source code can be found in CPG 7132a.l5,
Source Code for Process Control Application Programs, and include consideration of
the source code as master production and control records. Accordingly, those sections in
the regulations relevant to master production and control records are applied to the
computer application as source code. For custom-built applications, the program listings
are considered to be source code; for configurable applications, the configurable elements
or scripts are considered to be source code; and for off-the-shelf applications, the critical
algorithms, parameters, and macro listings are considered to be source code.
CPG 7132a.08, Identification of Persons on Batch Production and Control Records
issued in 1982, allows drug manufacturers to replace certain functions performed by
operators with computer systems. Part 211.101(d) requires verification by a second
person for components added to a batch. A single automated check is acceptable if it
provides at least the same assurance of freedom from errors as a double check. If it does
provide the same assurance, the process does not gain by applying a redundant second
check, which adds nothing to the assurance of product quality. The equivalency of an
automated single-check system to a manual check must be shown, however, and this
might not always be possible.
A major compliance issue is the administration of electronic records. Regulatory
authorities hold owners of records required by existing regulation responsible for
assuring the compliance of the recording and management of the electronic records. The
FDA establishes such responsibility in the CPG 7132a.12, Vendor Responsibility and
Part 11 has now established the detailed requirements for such records.
In 1983, the FDA provided another important guideline applicable to computer
hardware and software performing functions covered by the regulations. Their Guide to
Inspection of Computerized Systems in Drug Processing addresses the applicability of
the regulations to computer systems. According to this guideline, computer systems
hardware and software are considered equipment and records, respectively, within the
context of the regulations. One year later CPG 7132a.11,6 was issued, confirming the
applicability of the regulations to computer hardware and software. In the absence of
explicit regulations addressing computer systems, the regulations provide the implicit
guidelines necessary to meet the agencys expectations. In accordance with CPG
7132a.11, Table 31 lists the primary sections in the regulations applicable to computer
systems performing functions covered by the GMPs. Equivalent sections can be found in
the other FDA regulations.
6
FDA, CPG 7132a.11, Computerized Drug Processing, CGMP Applicability to Hardware and
Software, 9/4/87.
21 CFR Part 11
18
Table 31. US Drugs GMP.

US Drugs GMP Description
211.22
211.25
211.42
211.63
211.67
211.68
211.68
211.68(b)
211.68(b)
211.100
211.101(d)
211.105(b)
211.180
211.180(a)
211.180(c)
211.180(d)
211.182
211.188(a)
211.188(b)
211.189(e)
211.192
211.220(a)7
Responsibilities of QC Unit
Personnel qualifications
Design and construction
Equipment design, size, and location
Cleaning and maintenance
Maintenance and calibration
Written procedures
Record controls
Validation of computer systems (implicit requirement)
Written procedures, deviations
Double check on computer
Equipment identification
General (records and reports)
Records retention
Storage and record access
Records medium
Use of log(s)
Reproduction accuracy
Documentation and operational checks
Records review
QC record review
Validation of computer systems (explicit requirement)
1996 CGMP proposed regulations, FR, Vol. 61, No. 87, May 3, 1996.
19
Table 32. Reconciliation GMPs, EU Annex 11

and Part 11.
US Drugs
GMP/CPG
EU GMPs8
Description
Part
11
CPG 7132a.07
CPG 7132a.08
211.68(b)
211.220(a)*
211.180(a)
Annex 112
Validation
11.10(a)
Annex 1112,
1113
Annex 1113,
14, 15, 16
Annex 118
Annex 1110
Generation of accurate and complete

copies of records
Protection of records
11.10(b)
Annex 116
Use of operational system checks

to enforce permitted
sequencing of steps and events,
as appropriate
211.68(b)
211.68(b)**
Comment
paragraph 186,
1978 CGMP
revision
CPG 7132a.15
CPG 7132a.08
211.100;
211.188(b)
211.68(b)
n/a
211.25
211.180
n/a***
Subpart D
211.180(a)
Annex 1110
Annex 118
11.10(c)
Limiting access to authorized individuals 11.10(d)

Use of audit trails
11.10(e)
Authority checks
Use of device (e.g., terminal) checks
to determine, as appropriate, the
validity of the source of data input or
operational instruction
Annex 111
CVs and training records for those using
e-records
Annex 114
Controls over systems
documentation
Annex 118
Controls for identification
codes/password
Annex 113, 4, Controls for computer hardware
6, 13, 15
Annex 1114
Record retention
11.10(f)
11.10(g)
11.10(h)
11.10(i)
11.10(k)
11.300
11.10(d)
Not
covered
8
Wyn, Sion, Regulatory Requirements for Computerized Systems in Pharmaceutical
Manufacture, Software Engineering Journal, Vol. 11, No. 1, 1996, pp. 8894.
* Proposed changes to CGMP, May 1996.
** Sec. 211.68(b) requires appropriate controls over computer or related systems to
ensure that only authorized personnel make changes in master production and control
records or other records.
*** Implicit regulations can be found in 211.68.
21 CFR Part 11
20
The introduction in 1997 of Part 11 provided the formal codification applicable to

computer systems performing FDA-regulated operations. One of the fundamental
principles of Part 11 is that it requires organizations to store regulated electronic data in
its electronic form, once a record has been saved to durable media, rather than keeping
paper-based printouts of the data on file, as had been the long-term practice in
organizations performing regulated operations9. If information is not recorded to durable
media, the stored data will be lost and cannot be retrieved for future use.
When retrievability is an attribute, the procedural and technological controls
contained in Part 11 are essential to ensure record integrity. Only regulated electronic
records that meet Part 11 can be used to satisfy a predicate rule.
The current federal legislation embodied in the Electronic Signatures in Global and
National Commerce Act, defines electronic records in more general terms than Part 11.
The primary requirement of the current federal legislation is whether data is accessible
once it is put into storage, rather than the technology used. This definition considers data
in transient memory as an electronic record. Therefore, one key element to be analyzed
by both the FDA and the industry are the regulatory requirements for data that stored in
transient memory, including any audit trail information.
The controls on records contained in 21 CFR 11 Subpart B reconcile earlier GMPs and
policies. Table 32 associates the clauses in Subpart B 21 CFR Part 11 with the drugs
GMP regulations, applicable FDA CPGs, and the EU GMP regulations.
CPG 7153.17 is the most recent CPG applicable to computer systems. Published in
1999, the objective of this CPG is to establish the criteria that the FDA will take into
account when regulatory action is needed for a computer system found to be
noncompliant with Part 11. Refer to Chapter 21.
Another key concept in the FDA regulatory environments is the security of computer
resources. Chapter 14 covers this subject.
Regulatory authorities hold the owner of the records known to be required by existing
regulation responsible for assuring the compliance of the computer systems that record
and manage such records. The FDA defined this responsibility in CPG 7132a.12,
Vendor Responsibility.
MEDICAL DEVICES SOFTWARE

The primary FDA regulations applicable to medical device product software are 21 CFR
820.70(i), 21 CFR 820.30 and all of the CPGs discussed in this chapter. Validation
requirements apply to software used as components in medical devices, to software that is
itself a medical device, and to software used in production of the device or in
implementation of the device manufacturers quality system. According to 21 CFR
820.70(i) the automated system must be qualified for its intended use to ensure system
9
Note that Part 820 requires that results of acceptance activities be recorded but not necessarily
all raw data. Results must have audit trails. Be sensitive to need for raw data during failure
investigations under CAPA. Refer to Part 820 preamble, pp. 52631 and 52646.
21
performance. There must be an approved procedure and qualification protocol(s) in order

to ensure that qualification is properly performed. The FDA also requires that electronic
components of medical devices be regulated. An evaluation of component suppliers and
qualification of device manufacturing line(s) are elements of this qualification.
In order to meet the quality system regulations, electronic components are considered
to be similar to process equipment and components, and must be traceable from the
suppliers to the user. According to the FDA, software development is primarily a design
process. 21 CFR 820.30 contains the design control regulations applicable to medical
devices. The application of this section to product software10 includes all design activities
necessary to obtain, review, implement, and validate a design specification.
The design specification describes, in a narrative and/or pictorial form, how the
software will accomplish the software requirements and the interactions with the
hardware to accomplish various functions of the devices design.11 The most recent FDA
validation guidance,12 which is exclusively applicable to medical devices, provides
comprehensive information on the implementation of applicable software validation
regulations to medical device computer systems performing regulated operations.
In the context of Part 11, one of the differences between Parts 211 and Part 820 is the
scope of the electronic records that must be maintained. Part 21 requires the maintenance
of electronic regulated data including raw data. Part 820 requires that the results of
acceptance activities are recorded, but not necessarily all of the raw regulated data. The
exception in 820 is that raw data is required during failure investigations. As in Part 211,
results under Part 820 must have audit trails. Refer to Part 820 preamble, pp. 52631 and
52646.
THE FOOD INDUSTRY

FDAs authority to regulate the use of computers in food plants is derived from FD&C
Act Section 402 (a) (3). Specifically, conventional food is governed by the CGMP in
Manufacturing, Packaging, or Holding Human Food (21 CFR Part 110, April 2001). 21
CFR 110.40(d) provides the regulatory requirements applicable to computer systems:
Holding, conveying, and manufacturing systems, including gravimetric, pneumatic,
closed, and automated systems, shall be of a design and construction that enables them to
be maintained in an appropriate sanitary condition.
10
O.Lpez, Applying Design Controls to Software in the FDA-Regulated Environment, Journal of

cGMP Compliance, July 1997.
11
FDA, Application of the Medical Device GMP to Computerized Devices and Manufacturing
Processes, May 1992
12
FDA, General Principles of Software Validation; Final Guidance for Industry and FDA Staff,
January 2001.
21 CFR Part 11
22
There is no additional explicit citation to computer systems in Part 110. Taking into
account CPG 7132a.11, computer systems are referenced implicitly in multiple areas.
Part 11 covers electronic record requirements in Part 110. FDA regulates dietary
supplements under a different set of regulations than those covering conventional foods
and drug products (prescription and over-the-counter). Under the Dietary Supplement
Health and Education Act of 1994 (DSHEA), the dietary supplement manufacturer is
responsible for ensuring that a dietary supplement is safe before it is marketed. FDA is
responsible for acting against any unsafe dietary supplement product after it reaches the
market. At the present time, moves are being made to ensure that manufacturers of
dietary supplements register with FDA, or get FDA approval, before selling such
supplements. Manufacturers must make sure that product label information is truthful and
not misleading. FDAs post-marketing responsibilities include monitoring safety, e.g.,
voluntary dietary supplement adverse event reporting, and product information, such as
labelling, claims, package inserts, and accompanying literature. The Federal Trade
Commission regulates dietary supplement advertising. At this moment there is no
regulation covering the manufacturing of dietary supplement. Currently there is a
proposed ruling for the Current Good Manufacturing Practice in Manufacturing,
Packaging, or Holding Dietary Supplements.13 Under Equipment and Utensils, explicit
reference on computer systems performing functions in FDA-regulated products:
(a) (b) Holding, conveying, and manufacturing systems, including gravimetric,
pneumatic, closed, and automated systems, shall be of a design and construction that
enables them to be maintained in an appropriate clean condition
indicates that computer-related regulatory provisions for the food industry would be
equally applicable to the dietary supplements industry.
13
FR, Vol. 62, No. 25, February 6, 1997.
Chapter 4
New Computer Systems Validation Model1
In the past, the approach to computer qualification or validation and associated
configuration management in the regulated industries was based on key practices. The
foundations of these key practices are contained on publications such as:
Guideline on General Principles of Process Validation, May 1978
Compliance Policy Guidelines, September 1982
Guide to Inspection of Computerized Systems in Drug Processing, February 1983
PMA Staying Current Series (series started in May 1986)
Application of the Medical Device GMPs to Computerized Devices and Manufacturing
Processes, November 1990
Good Automated Manufacturing Practices (GAMP), February 1994
PDAs Validation of Computer Related Systems, October 1994
General Principles of Software Validation; Final Guidance for Industry and FDA Staff,
June 2001
PDAs A Globally Harmonized Glossary of Terms for Communicating Computer
Validation Key Practices, PDA Journal of Pharmaceutical Science and Technology,
Vol. 53, No. 2, March/April 1999
Best Practices for Computerized Systems in Regulated GxP Environments, PIC, Draft
Rev. 3.01
The constraint of key practices is that they need to be monitored and evaluated
periodically to ensure that they are suitable, and to keep them current with industry and
regulatory trends.2 Across the industry, key practices were not always implemented at
the same time, causing divergence in the implementation of the related key practices.
Part 11 provides the explicit and current regulatory trends applicable to computer
systems performing regulated operations. Now that the regulatory expectations are
clearly established, the implementation of Part 11 is contingent on the availability of
appropriate applications, which are compliant with the regulation. In the absence of
technologies supporting Part 11, the required controls are procedural in nature.
Figure 41 depicts a 21 CFR Part 11 Model developed by the author of this book1
helps determine Part 11 requirements or consideration during the development and
qualification of computer systems based on three (3) main characteristics:
1
Lpez, O., FDA Regulations of Computer Systems in Drug Manufacturing13 Years Later,
Pharmaceutical Engineering, Vol. 21, No. 3, May/June 2001.
2
Grigonis, Subak, and Wyrick, Validation Key Practices for Computer Systems Used in Regulated
Operations, Pharmaceutical Technology, June 1997.
New computer systems validation model
25
Figure 41. 21 CFR Part 11 Model.

whether the computer system is closed or open
whether the system creates electronic records
whether the system uses electronic signatures
In general, the elements to be verified and/or qualified during software implementation
are
the type of system (open or closed)
security functions
audit trails
operations controlled by the computer systems (operational checks)
implementation of the technology to support the process and Part 11
21 CFR Part 11
26
These two latter elements are not identical on all systems. Electronic signatures are tested
when the technology is implemented. An element out of the scope of this model is the
retention of electronic records, but the model should be used to verify and validate the
implementation of the system(s) that will hold these records. For applications using
electronic signatures, the current validation practices requires inspections and/or testing
for many of the following technical controls:
1. Open/Closed systems
2. Security
a. System security
b. Electronic signature security
c. Identification code and password maintenance
Identification code and password security
Passwords assignment
d. Document controls
e. Authority, operational, and location checks
f. Records protection
3. Operational Checks
4. Audit Trails
a. Audit mechanism
b. Metadata
c. Display and reporting
5. Electronic signatures
a. E-sign without biometric/behavioral identification
b. E-sign with biometric/ behavioral identification
c. Signature manifestation
d. Signature purpose
e. Signature binding
6. Certification to FDA
A subset of the above is applicable for hybrid computer systems.3 The implementation of
these requirements has been discussed in other articles.4,5
3
In hybrid systems, some portions of a record are paper and some electronic.
Lpez, O, Automated Process Control Systems Verification and Validation, Pharmaceutical
Technology, September 1997.
5
Lpez, O, Implementing Software Applications Compliant with 21 CFR Part 11, Pharmaceutical
Technology, March 2000.
4
New computer systems validation model
27
Figure 42 depicts the progression of the computer systems validation practices, from
the current best practices to the 21 CFR Part 11 Model. The transition program consists
of educating the regulated industry, assessing current computer systems, and
implementing the regulations using the appropriate technologies.
Figure 42. Progression of

Computer Systems Validation Key
Practices6
This model can be easily applied to the new generation of computer systems by
incorporating the requirements contained in Part 11 at the beginning the development
process. For the current generation of systems, an assessment will have to performed to
evaluate the level of conformity with the regulation.
Examples of the applicability of this model can be found in Appendix C.
A Partnership Approach to Achieving Regulatory Compliance for Electronic Records and

Signatures, paper presented at the October 1999 IMPACC Conference.
Chapter 5
Computer Validation Management Cycle
The implementation or update of a computer system is planned and executed in
accordance with the organizations own project management practices. It is the project
managers responsibility to take into account the technological requirements, standards,
procedural controls, regulatory requirements and related guidelines, and industry
standards.
During the implementation or update process, the introduction of a computer
validation management cycle will provide a top-down approach to integrate computer
systems validation strategies in order to support the project. Figure 51 suggests a
computer validation management cycle.
Figure 51. Computer Validation

Management Cycle.
As with any process, an improvement mechanism is required in order to enhance the
computer validation management cycle. The improvement mechanism will be identified
based on experience in using the SLC, validation practices, and the analysis of software
metrics. The following sections describe the overall scope of each document or activity.
Detailed description of each activity can be found in other chapter of this book.
Computer validation management cycle
29
VALIDATION POLICIES1
Policies define the general principles and philosophy that are required within an
organization. Each company should have written and approved validation policies that
communicate the expectations of senior management for the execution of a validation
project.
Company name is committed to ensuring that the products, equipment, facilities,
computerized systems, and support systems directly involved in the manufacture,
testing, control, packaging, holding and distribution of marketed products are
compliant with cGMP regulations. This commitment extends to the manufacturing,
facilities, utilities, processing, cleaning, computerized systems, analytical and
microbiological methods, packaging and support systems involved in the development
and production of pharmaceutical products subject to cGMP regulations. To attain
this level of compliance, management will ensure that proper validation and
qualification will be carried out at all stages of the product development cycle.
Any new equipment, process, product, facility or system will be prospectively
validated prior to becoming operational in the manufacture of finished dosage forms.
Concurrent validation and qualification will be considered for any existing
equipment, facility, product or system where limited documentation of appropriate
technology background exists, and an acceptable level of confidence can be
established by reviewing and documenting the operating history. Concurrent
validation will be employed where applicable, to certify that an appropriate level of
process control exists for all pertinent products, equipment, facilities and systems.
Retrospective validation or qualification is discouraged, but may be applicable in
some situations, where a statistically significant number of batches with consistent
formulas, procedures and analytical methods are available. Sufficiently detailed past
processing and control records must be available for retrospective validation studies
to be considered.
Sidebar 51. Sample Policy.
Policies are intended to establish responsibilities and expectations. All computer systems
validation procedures, guidelines, and strategies must be consistent with the Validation
Policy.
An example of a corporate-wide validation policy is contained in Sidebar 51.
In the context of Part 11, it is expected that information shall be characterized and
information quality incorporated into the computer systems development and operational
life periods.
1
PDA Committee on Validation of Computer-Related Systems, Technical Report No. 18

Validation of Computer-Related Systems, PDA Journal of Pharmaceutical Science and
Technology, Vol. 49, No. S1, 1995.
21 CFR Part 11
30
VALIDATION GUIDELINES
Validation guidelines must comply with all formal procedural controls established by the
company/organization. They are internal documents providing detailed procedures,
activities and sample products necessary to successfully execute a computer systems or
related project.
Typically, a guideline provides recommendations, not requirements, and represents
current thinking on the topic covered within the guideline. An alternative approach may
be used, however, if such an approach satisfies the requirements of the applicable
procedural controls and predicate regulation. It is often the practice of the FDA-regulated
industries to not refer to guidelines in SLC related documentation. Guidelines and
operating procedures may be contained in a Quality Plan, model specified in ISO
0005:1995, which defines and documents how the requirements for quality will be met
for a project. Quality Plans and Validation Project Plans complement each other.
VALIDATION PLANS
Refer to Chapter 8.
PROCEDURAL CONTROLS
Refer to Chapter 12.
COMPLIANCE ASSESSMENTS
Compliance assessments for computer systems are performed periodically, based on the
applicable predicate regulatory and Part 11 requirements. These assessments must be
performed in order to identify any functional gaps, and/ or procedural gaps, which may
be present for each computer system implemented. The analysis will determine if
operational, maintenance or security controls, specific to the system, provide a controlled
environment ensuring the integrity of the electronic records and/or signatures as stated in
the regulatory requirements. Additional information can be found in Chapter 19.
VALIDATION OF COMPUTER SYSTEMS

Refer to Chapter 2.
SUPPLIER QUALIFICATION
The qualification of computer technology suppliers and contract developers is an ongoing
activity. The objective of this process is to evaluate the computing environment and
Computer validation management cycle
31
technology products of the supplier, against established quality standards, managing any
noted deficiencies that are discovered, and working with the supplier to improve the
quality of its products, services, and documentation. Chapter 17 provides information on
the supplier qualification activities.
ONGOING SUPPORT SYSTEMS

After the system has been released for operation, system maintenance activities take over.
The importance of such activities is characterized by recent FDA remarks related to the
lack of change control management by regulated organizations. The FDAs analysis of
3140 medical device recalls conducted between 1992 and 1998 reveals that 242 of them
(7.7%) are attributable to software failures. Of those software related recalls, 192 (or
79%) were caused by software defects introduced when changes were made to the
software, after its initial production and distribution.
The maintenance activities must be governed by the same procedures followed during
the Development Period.
Chapter 18 provides information on ongoing support systems.
Chapter 6
Computer Validation Program
Organization
Many groups, both within and outside the organization, support the validation of
computer systems. It is the responsibility of Executive Management to provide adequate
resources to support the achievement of compliance in the computer systems validation
area.
Responsibilities are often similar across multiple functional groups such as
Information Technology, Engineering, Laboratories (QA/R&D), Clinical Research,
Manufacturing/ Operations, contract developers, and computer technology suppliers.
In order to manage the execution and supervision of computer systems validation
activities, there needs to be an organizational structure established. This chapter suggests
an organizational structure for supporting to a computer validation program.
ORGANIZATIONAL MODEL
This model (Figure 61) is an example of how computer systems validation and its
related activities can be organized and responsibilities allocated. Each organization
should prepare a document to identify the roles and responsibilities appropriate for its
business environment.
Figure 61. Organizational Model.
Computer validation program organization
33
COMPUTER SYSTEMS VALIDATION EXECUTIVE

COMMITTEE
The Computer Systems Validation (CSV) Executive Committee is composed of
Validation, Compliance, Regulatory Affairs, and Information Technology senior
management. Another possible member of this committee is a Process Innovator.
The CSV Executive Committee is responsible for:
reviewing and approving computer systems validation guidelines
reviewing and approving the computer systems validation master plans
providing direction and guidance concerning the validation program
arbitrating on organizational issues and validation program issues
supporting the awareness and implementation of the CSV strategy through resourcing
training, and procedure and guideline development
CSV CROSS-FUNCTIONAL TEAM

The Cross-Functional Computer Systems Validation Team is a multidisciplinary team
(Figure 62) that represents the overall needs of the business. This team comprises
representatives from research production, engineering, quality control, clinical research,
quality assurance, and other groups as needed.
Figure 62. Cross-Functional Team.

The members and their deputies are assigned by the CSV Executive Committee. For ad
hoc activities, the team can be complemented with representatives from departments
performing activities that support computer validation. Examples of such departments
are: Purchasing, IT, QA Validation, QA Compliance, Operations, R&D, Engineering, and
representatives from each manufacturing facility concerned.
21 CFR Part 11
34
The Cross-Functional CSV Team is responsible for:

reviewing and approving the CSV procedural controls
organizing the implementation of the validation policy, validation guidelines, and
computer system validations procedural controls
developing and maintaining the CSV approach to be followed by the validation team
reviewing validation plans and protocols as needed
assisting the validation program coordinator in compiling the validation master plans
developing and maintaining a QA framework for CSV activities
assisting QA Compliance during audit of a computer system
support computer-related company and/or executive committee strategies
The importance of the QA unit in validation is evident from the regulations. The QA unit
has the responsibility and authority for:
reviewing and approving the validation policy, validation guidelines and validation
procedural controls
reviewing the computer systems validation work and documentation and signing to
verify the correct execution of the validation plan, protocols, and summary report for
computer systems performing regulated operations
assessing the compliance of computerized systems to the validation policy, validation
guidelines, validation procedural controls, and to related predicate regulations
auditing the readiness of computer systems for a regulatory inspections
supporting Line Management Groups during regulatory inspections
documenting the results of audits in QA audit reports
ensuring that any audit report issues are resolved
organizing capability assessments on external and internal suppliers
CSV GROUPS AND TEAMS

The CSV Groups and Teams mainly focus on the implementation of computer systems
validation projects.
The CSV Group is responsible for the delivery of a computer validated system. This
includes:
Performing the validation work for GxP computer systems from the plan through to the
report and producing the required documentation
Ensuring that the computer system is maintained in a validated state by ongoing
qualification or validation of the platform
The member of the CSV Group include the:
System Validation Team designated by the Management Group
Supplier of the Standard/Configurable Software Packages used and/ or the System
Integrator (software system and platform)
Platform Owner
35
The CSV Team includes persons in the organization who plan and execute the CSV
activities, including the System Owner, System Users, Validation Coordinator, and
specialists in various fields related to computer systems validation. The CSV Team is
responsible for:
developing, approving and executing the CSV Plan
developing and approving protocols and qualifications reports and
writing the Validation Summary Report
One of the most important members of the CSV team is the Validation Project
Coordinator. Responsible for planning, coordinating, and reporting the validation
activities throughout a project, the validation coordinator may act as the liaison between
the Validation Program Coordinator, other teams and groups in the organization. The
Validation Project Coordinator will act as a consultant and must be familiar with the
contents of the validation practices and procedures; and must have participated in
appropriate training activities in the area of computerized systems validation.
When embarking on the implementation of a regulatory computerized system, it is
recommended that the project managers appoint an individual to act as the validation
project coordinator. The involvement of the validation coordinator on a project will vary
depending on the size of project. For a small project, the validation coordinator role may
require only a part-time assignment. For a large project, the validation coordinator role
may require the assignment of a full-time person.
THE MANAGEMENT GROUP

The Management Group consists of the:
System Owner
System User
The Management Group is responsible for:
preparing, approving, and implementing procedural controls that are appropriate for the
system
managing the system throughout its entire project life cycle
preparing the information needed by the Validation Project Coordinator for the
computer validation master plan
setting up, managing, and resourcing the CSV Team
approving the Validation Reports for the computerized systems
archiving the system validation documentation
ensuring the validation status is maintained throughout the operational life of the
systems
providing support during regulatory inspection
21 CFR Part 11
36
VALIDATION PROGRAM COORDINATORS

This is a dedicated corporate consulting group. The principal task of this group is to
provide consultation and to approve validation documentation; a benefit of this approach
is that a consistent computer validation program is implemented across all sites and units.
For any validation activity to be effective, an independent and qualified third party
(someone who is neither a developer nor a user of the system) must review it. Although
the developers and usersof the system may generate most of the SLC products, a
qualified third party should review thecomputer systems validation records and give
objective assurance to management that thecomputer systems validation was properly
carried out. This third-party review function can beperformed by a dedicated validation
group, with sufficient objectivity and expertise in computersystems testing and validation
methods. Members of a CSV group should include system,software, and hardware
engineers with experience in regulations and project management; adedicated CSV group
needs experience and knowledge of:
computing infrastructures
security and controls
corporate, industry, and regulatory standards
solving computer systems problems
the roles and responsibilities of the integrators
The Validation Programs Coordinators may be part of the QA unit, Technical Services
unit, or the Information Technology unit.
The advantages of the dedicated Validation Program Coordinators are that employees
in such a group are totally dedicated to (and responsible for) the computer validation
effort. Interaction with production and quality assurance scheduling is important and
must be taken into consideration.
The value added by a dedicated CSV group to the validation process includes:
reducing the learning curve to the Management Group, System Validation Team, and
System Validation Group by providing validation consultation
producing a smoother path for reviewing the manufacturing systems validation
documentation and reviewing protocols at the start of the development effort
reducing inefficiencies by ensuring the quality and consistency of the validation
methodology
improving working inefficiencies by reusing validation experience between similar
computer systems and in manufacturing facilities
To summarize, the Validation Program Coordinators are responsible for:
managing the CSV program, consisting of the Validation Policy, validation guidelines,
validation procedural controls development and implementation, computer validation
training, and other activities that have been identified
37
coordinating the development of validation guidelines and validation procedures and

organizing regular reviews, and the development and review of generic validation
plans and protocols
reviewing computer validation work and documentation and signing to verify the
correct execution of the validation plan, protocols and summary report for computer
systems and to ensure the uniformity and quality of the approach
compiling the Computer Validation Master Plan
keeping an inventory of the validation status of the computer systems
providing support, consultancy, tools (e.g., templates) and expertise on CSV matters to
the CSV teams
supporting QA during audits that assess the readiness of computer systems for
regulatory inspections
supporting Line Management Groups during regulatory inspections
developing and providing validation training based on procedural controls
Chapter 7
The Computer Systems Validation Process
It is an FDA requirement that regulated organizations ensure that all computer systems
performing functions under FDA predicate regulations conform to the requirements
contained in Chapter 3. In the FDA regulation context, computer systems validation is an
element of the SLC. The scope includes all the practices and tools necessary to manage
the development and maintenance of the deliverable system.
All new computer systems must be prospectively validated before going into
production. For existing computer systems, concurrent validation may be considered
where limited documentation on the appropriate technology exists, and where an
acceptable level of confidence can be established by reviewing and documenting the
operating history.
For existing computer systems, retrospective evaluation is discouraged. It is
extremely difficult to evaluate a computer system retrospectively, being generally more
costly and time consuming than prospective validation. Retrospective evaluation should
be used only as a corrective measure in response to deficiencies noted concerning prior
validation efforts. See Chapter 19 for a brief discussion on this subject.
Any deviations raised during the application of a computer system validation
procedure shall be documented following the applicable deviation recording procedure.
The following describes the process for developing and maintaining computer systems
in an FDA context.
The depth of validation and the steps of the validation process to be followed will vary
depending on the category of software (e.g. as the one described in GAMP)1 and the
complexity and criticality of the system.
The depth of the required steps is estimated using the Criticality and Complexity
Assessment and may be performed during the early Project Conceptualization stages of a
project; and reevaluated at Project Initiation. A sample Criticality and Complexity
Assessment can be found in Appendix D.
Each phase of the SLC must be controlled to maximize the probability that a finished
system meets all quality, regulatory, safety, and specification requirements. If an SLC
approach is applied properly, no additional work will be required to validate a system.
For each SLC period and event, computer systems validation requires that the
development processes are documented work products. As explained in Chapter 2, phase
gate verification activities performed during each event may be a perfect place to review
and quantify the quality of all products needed to support the next phase.
1
GAMP Guide for Validation of Automated System, Version 4.0, ISPE (GAMP Forum), December
2001.
The computer systems validation process
39
The elements to be included as part of the computer systems validation process will vary depending
on the category of the software. It is the practice in the FDA-regulated industry to take into account
five (5) categories of software.
Figure 71. Conventional Waterfall

Model.
2
K.Baker and T.Simmons, Software Management Basics (ISBN 0954979607), Sue Horwood
Publishing Ltd., Storrington, U.K.
Operating systems (e.g., Windows)

Standard instruments (e.g., bar code readers, some laboratory instruments)
Standard software packages (e.g., spreadsheets, databases)
Configurable software packages, (e.g., LIMS, SCADA, building automation systems)
Custom-built systems (e.g., automated process control systems)
Chapter 10 provides a comprehensive description of the validation elements to be
considered for each software category.
Figure 23 illustrates a typical SLC. Using as an example, a high-criticality and
highcomplexity, custom-built system, the conventional V model (Figure 71) as
framework, and the typical waterfall methodology as development methodology, the
following sections summarize the typical project activities.
Detailed descriptions by SLC Periods can be found in Appendix E.
21 CFR Part 11
40
1st PeriodConceptualization
During this period the computer system high-level functionality is agreed upon with the
users, system owner, developers, and quality for the operation and automation of the
operation. Part 11 considerations are a key element in this phase, which includes the
following activities.
Initialization of the project to automate or to re-engineer an operation
Review project proposal
Develop operation study
Perform preliminary evaluation of the technology available
Refine scope of the system and confirm strategic objectives
Develop a process/operating description, associated restrictions, interfaces,
proposed solution, and critical operational checks/sequencing, including how
Part 11 will affect and interact with the process/operation
Approve the operational feasibility study
Perform a preliminary Criticality and Complexity analysis
Business and regulatory decision to continue or to hold the project
Appendix E contains a comprehensive description of the Conceptualization Period.
2nd PeriodDevelopment
The purpose of the Development Period is to specify, design, build, test, install, and
support the business operation to be automated or updated, and approved in the
Conceptualization Period using a structured process. Part 11 must be taken into account
during all of these activities. The development process involves the following steps:
Requirements gathering
Interviews, surveys, studies, prototypes, demonstrations, and analysis methods
may be used to gather requirements
Gather the requirements for the systems including functional (e.g. operational
checks) requirements, nonfunctional (e.g., coding standards) requirements,
users, company-wide regulatory compliance (e.g., Part 11 technical control),
safety, process, and other applicable requirements
Characterize information, assess its value to the organization, and incorporate
information quality as part of the project plan
Conduct a system (hardware, software, and process) risk analysis. New
requirements may be found as the result of the risk analysis. Any new
requirements must be documented in the requirements specification
deliverable
Data definition, database requirements, and records storage requirements
Develop a requirements specification deliverable
Review the requirements specification deliverable against the conceptual
system requirements established during the Conceptualization Period. Any
discrepancy must be resolved and the documentation amended
41
Develop a validation plan

Complete a risk analysis and a criticality/complexity analysis
Complete change control and approvals
System specification
The system specification deliverable describes in detail what the system should
do, but not how (in terms of technology) it should be done. It is based on the
requirement specification deliverable and the suppliers solution.
Approved requirements specification deliverable
If the software will not be developed in house, conduct a software supplier
audit
If elements of the hardware are unique, conduct a hardware supplier audit
Select supplier(s)
Gather all of the measurable operational parameters and limits for the system
Develop high-level flow diagrams and general screen layouts
Review the system specification deliverable. The objective of this review is to
verify that the all of the specifications are correct, consistent, complete,
unambiguous, testable, modifiable, and traceable to the concept phase
documentation
Perform and document a traceability analysis between the requirements
specification deliverable and system specification deliverable
Specify Part 11 technical controls
Prepare and approve system specification deliverable
Revisit the risk analysis and the criticality and complexity assessments based
on the system specification deliverable
Review validation plan based on the system specification deliverable
Approved validation plan
Begin to plan the test activities, including unit testing, integration testing,
factory acceptance testing, site acceptance testing, and system qualifications
Technical design
System requirements are applied to the software design. During the technical
design phase, each requirement described in the system specification
deliverable is implemented. Also, subsystem components and interfaces, data
structures, design constraints, and algorithms are developed as part of the
system design decomposition process. This activity is very critical for medical
device companies. The design inputs are contained in the requirements for the
computer system and the design output(s) can be provided as part of the design
specification or the deliverable for the activity concerned3.
3
O.Lpez, Applying Design Controls to Software in the FDA-Regulated Environment, Journal of

EGMP Compliance, Vol. 1, No. 4, July 1997.
Design according to SLC procedures

computer hardware and software architecture
data structures
21 CFR Part 11
42
flow of information. This design phase forms the input for the
development of the integration test and operational checks
interfaces
Assemble/collate the design deliverables
Perform design reviews. Verify if the risks that were previously identified
were mitigated as part of the solution presented in the design
Finalize the test planning
Design the Part 11 technical controls
Approve the technical design specification deliverable(s)
Conduct in-process audit activities associated with the technical design
process
Revisit the risk analysis
Begin planning the development of procedural controls for those Part 11
requirements that are not covered by technology
Program build
Provide the technical design product(s) to the programmers
Translate the design to a target program
Develop programs based on the development (e.g., coding) standards, and
build Part 11 technical controls into the system
Review code integrity using static analysis, audit(s), inspection(s), review(s),
and/or walkthrough(s)
Conduct software unit testing. The machine executable form of the code is
tested to uncover errors in function, logic, or implementation
Integrate modules and perform integration testing
Test the Part 11 technical controls
Preparation of the qualification protocols4
An IQ is required if:
new software/hardware is being installed
existing software/hardware has been modified or relocated
existing software/hardware was never qualified
An OQ is required if:
operating parameters and sequences (e.g., operational checks) are being
qualified for new software/hardware
sequences have been modified
new operating parameters are being established for existing
software/hardware
existing software/hardware was never qualified In some environments,
an OQ is required to test software in testing environment.
A PQ is required if:
PQ is always required to test software in the production environment
43
Authors note: While IQ/OQ/PQ terminology has served its purpose well and is one of many
legitimate ways to organize dynamic testing at the operational environment in the FDA-regulated
industries, this terminology may not be well understood by many software professionals. However,
organizations performing regulated operations must be aware of these differences in terminology as
they ask for and provide information regarding computer systems.
the new software/hardware or changes can directly affect the

management of electronic records
existing hardware/software was never qualified In some environments,
PQ is required, if after performing the OQ in the testing environment,
testing software is performed again in preproduction environment
Conduct an FAT
Conduct an SAT
Conduct an IQ
Conduct an OQ
Under certain conditions, an FAT may replace an IQ and/or OQ. Refer to
Appendix E.
Conduct a PQ
Develop qualification report(s) and/or a validation report
Test results and data are formally evaluated
Conduct a Process/Product PQ. Note: typically, this activity is performed on
manufacturing systems only. The objective is to demonstrate that the process
consistently produces a result or product meeting predetermined specifications
Prepare and approve a Product/Process Summary Report
Update any documentation that has references to the application software and
its associated version
The approved summary report and associated validation documentation are
submitted for retention to the National Quality Assurance Document Control
Center or site documentation retention center
3rd, 4th, and 5th PeriodsEarly Operational Life, Maturity, and Aging
hardware and software operation and maintenance
ongoing reviews/verification program
6th PeriodRetirement
elimination or replacement of computer systems
SYSTEM DEVELOPMENT FILES

One key element necessary to support the SLC is the availability and maintenance of the
system development files. The developer must document the development of each system
21 CFR Part 11
44
unit, system component, and configuration item in the software development files and/or
use a suitable configuration management tool.
The developer should establish a separate system development file for each unit or
logically related group of units; document and implement procedures for establishing and
maintaining system development files; and maintain system development files until the
system is retired. The system development files should be available for the agency to
review upon request; and be generated, maintained, and controlled by automated means.
To reduce duplication, each system development file should avoid containing information
provided in other documents, or in other system development files. The set of system
development files should include (directly or by reference) the following information:
design considerations and constraints
design documentation and data
schedule and status information
test requirements and responsibilities
verification and test procedures, and results
Chapter 8
Validation Project Plans and Schedules
REGULATORY GUIDANCE
The validation plan is a strategic document that should state what is to be done, the
scope of approach, the schedule of validation activities, and tasks to be performed. The
plan should also state who is responsible for performing each validation activity. The
plan should be reviewed and approved by designated management.1
VALIDATION PROJECT PLANS

Validation plans are documents that tailor a companys overall philosophies, intentions,
and strategy to establish performance and computer systems or software adequacy.
Validation plans state who is responsible for performing development and validation
activities, who identifies which systems are subject to validation, who defines the nature
and extent of inspection and testing expected for each system, and who outlines the
framework to be followed to accomplish the validation.
In general, validation project plans describe the organization, activities, and tasks
involved in the development of a computer system, including:
organizational structure of the computerization project
the departments and/or individuals responsible
resource availability
risk management
time restrictions
the SLC and development methodology to be followed
deliverable items
overall acceptance criteria
development schedule and timeline
system release sign-off process
sample formats for key documentation
Validation project plans and schedules
47
The plan may also refer to procedures to guide operation and maintenance.
Validation plans ensure that mechanisms are in place to guide and control the multiple
development or maintenance activities, many of which run in parallel, and assure proper
1
Signatures: Validation, August 2001.
communications and documentation. During the execution of a project, verification will

check the reliability of the work as established by the project plan, so that when the due
dates for completion/hand-over arrive, a high degree of certainty exists that the system is
validated.
When multiple departments are involved in a project, the system owner will take
responsibility for the validation documentation. Other departments will provide
documentation and personnel to support the development, validation, and maintenance
effort.
Validation plans are not required by any of the predicated regulations but are
considered a key project management practice. Validation plans are an essential
document for the overall management of projects, and are crucial for the success of the
projects.
Typically, managers, their peers, end-users, and those responsible for delivering the
system, approve validation plans. Quality assurance may also sign the document. The
validation project plan and the requirements specification deliverable, together define the
technical and regulatory requirements applicable for a project.
Project validation plans should be started during the early stages of a project. Initial
project concepts and planning estimates should be elements in the creation of a project
validation plans. The initial project verification activities will assess the project teams
capability to produce a validated system and provide input for defining the level of
testing effort expected. Project verification will identify any critical deviations to the
expected project timing and quality levels, as well as other issues affecting the timely
approval of the validation report.
An approved version of the validation plan should be available when a computer
technology supplier or contract developer is being selected, and should be updated
whenever project events or verification results require a change. An example of a project
event would be a change in project scope.
Documents that support the update of the validation plans are:
system requirements
criticality and complexity analysis
project verification results
other system descriptions
The format of validation project plans is flexible and may incorporate Gantt charts. The
contents of the validation plan may include, but are not limited to, the following list:
document control section
system/installation name
author(s)
21 CFR Part 11
48
creation, save and print date

version number
document identification
document history
reviewer and the review date
external document references
table of contents
intended audience
scope
objective
system description
validation acceptance criteria
verification activities and deliverables
qualification activities and deliverables
qualification planning
project verification
installation qualification
operational qualification
performance qualification
process validation
roles and responsibilities
application owner
project management
QA
validation team
computer technology supplier or contract developer
document control
project schedule
project documentation and its delivery
project activities
MANDATORY SIGNATURES
Validation plans must be reviewed and approved by the author, the organization owning
the execution of the plan, and by the system owner, because they will take the
responsibility for the plan.
49
Validation Plan and Protocols: One document? Many documents?

When developing the validation plan, a decision has to be made about the manner in
which the documents will be organised:
For small systems, it is possible to integrate the validation plan and all protocols
into one document, and to have one report to summarising all results.
For large systems with many components, a validation project plan can be created
that divides the validation effort into smaller, more manageable units with
separate plans.
Multiple versions of the plan and protocols could be needed, with interim or partial
reports.
The guiding factor for organising validation documentation is the criticality and
complexity of the system, and the type of software. Refer to Appendix D.
The review and approval of the plan by QA is optional, but will provide an endorsement
that the plan conforms to the current written procedures on computer systems validation,
and that the document incorporates applicable regulatory requirements.
PROJECT SCHEDULE
1. Validation plans should include a high-level project schedule that includes the
identification of resources. As further project information is known, additional
elements can be incorporated to the validation and project schedule process. Figure 8
1 2 depicts a typical systems development distribution showing the percentage of effort
needed for each key developmental activity.
Figure 81. Systems Development

Distribution.
21 CFR Part 11
50
The following list includes examples of items that need to be tracked as part of the
project schedule:
project kickoff
validation plan preparation, review, and approval
training of the validation team
applicable SLC periods and events
review and/or development policies, guidelines, standards, and procedural controls
quality checkpoints for reviews
project documentation due dates/reviews/approvals
qualify computer technologies suppliers and/or contract developers (if applicable)
development and approval of procedural controls
write specific plans and protocols for IQ, OQ, PQ
develop the traceability analysis and verify the traceability of the design and testing
elements against user requirements
review/approve specific protocols
monitor the execution of specific protocols
assemble, review, and approve the results
assemble, review, and approve the final validation report
2
Roger S.Pressman, Software Engineering: A Practitioners Approach, 4th ed., McGraw-Hill, NY.
Figure 82. Software Development

Stages Relative Costs.
Validation plans and associated schedules are live documents that should be reviewed
periodically. Phase gate verification activities, performed during each event, may be a
perfect place to review the project plan and schedule.
51
As part of the conceptualization period, senior managers are presented with, and
estimate the total effort to be expended (including maintenance for the project). The
relative costs of each development stage are depicted in Figure 82.2 Note: Maintenance
of existing software can account for over 70% of all effort expended by a software
organization and this cost will be passed on to the customer.
As part of the schedule, it should be considered that the majority of the development
of computer systems fails due to poor gathering of requirements. This situation leads to
negatively affect subsequent development activities and associated work products. An
investment up front can yield notable savings (cost avoidance) later.
Chapter 9
Inspections and Testing
REGULATORY GUIDANCE1
Structural Testing
Structural testing takes into account the internal mechanism (structure) of a system or
component. It is sometimes referred to as white box testing. Structural testing should
show that the software creator followed contemporary quality standards (e.g., consensus
standards from national and international standards development organizations. This
testing usually includes inspection (or walk-through) of the program code and
development documents.
Static Verifications
These include static analyses such as document and code inspections, walk-through, and
technical reviews.
Functional Testing
This testing involves running the program under known conditions with defined inputs,
and documented outcomes that can be compared to predefined expectations. Functional
testing is sometimes called black box testing.
21 CFR Part 11
54
Program Build Testing

Program build testing is performed on units of code (modules), integrated units of code,
and the program as a whole.
INTRODUCTION
Software engineering practices include documented unit testing; code reviews; explicit
highlevel and low-level design documents; explicit requirements and functional
specifications; structure charts and data flow diagrams; function-point analysis; defect
and resolution tracking; configuration management; and a documented software
development process.
The FDA expects the same quality principles to be applied during the development
and maintenance of computer systems; quality principles must be contained in procedural
controls.
Numerous inspection steps undertaken throughout the system development and
operational life support that a computer system is validated. These include static analyses
such as document and code inspections, walk-through, and technical reviews. These
activities and their outcomes help to reduce the amount of system level functional testing
needed at the operational environment to confirm that software meets requirements and
intended uses.
DOCUMENT INSPECTIONS AND TECHNICAL REVIEWS

Inspections are testing practices in which source code and documentation are examined in
either a formal or informal manner in order to discover errors, violations of standards,
inconsistencies, and other problems. The objectives of these reviews are to verify the
following:
consistency between the requirements and the specification at each level
numerical factors
safety and reliability
correctness of design
correctness of sequence
correctness of formulae and calculations
traceability to previous design and development levels (i.e., specifications)
testability
maintainability
usability of the product/documentation
that programming standards, including code layout rules, commenting and naming
conventions are consistently applied
that the code has been produced in accordance with the technical design specification
deliverable
Inspections and testing
55
During the development process a set of inspections (i.e., in-process audits) are
performed in order to verify the consistency of the code against the design work
products. These inprocess audits can be used to determine the following:
the efficiency of the design algorithms and code produced may be evaluated
the consistency of hardware and software interfaces
a review of the system requirements and design specification coverage for static and
dynamic testing (the software validation plan is the place to specify testing
activities)
consistency between the system requirements and the technical design
consistency between the technical design and the system specification
consistency between the code and the technical design description
After technical design has been translated into code,2 the following inspections should be
considered to further evaluate the correctness of the code. A code review is an informal
white box test where the code is presented in a meeting to project personnel, managers,
users, customers, or other interested parties for comment or informal approval. Code
reviews are a static test approach that can be applied to source code during the early stage
of development.
Code reviews are performed during software coding and are also applicable during the
Operational Life of the application software. They establish that the software can provide
all of the features specified during the technical design, with all of the necessary quality
attributes, before starting the dynamic testing (unit, integration, and/or black box). It
includes:
verifying the absence of dead code (CPG 7132a.15)
confirmation of a modular design
conformance with the required programming approach
verifying any design mitigation which have resulted from the risk analysis
verifying that the elements contained in the code will efficiently enable code to be
maintained during the its Operational Life
Code inspections are formal testing techniques that focus on error detection. The
programmer reads the source code, statement by statement, to a group who ask questions
that analyze the program logic, analyze the code against a checklist of historically
common programming errors, and verify that the code complies with the appropriate
coding standards.
Code audits are formal and independent reviews of source code by a person, a team, or
tool, to verify its compliance with software design documentation and programming
standards. In addition, the software code is evaluated against its supporting
documentation to ensure that it correctly describes the code.
2
Code or source code (refer to Chapter 16) must be viewed, in this context, in a wide meaning,
since it includes, for example, the program listings for custom-built applications; the configurable
elements or scripts for configurable applications; or the critical algorithms, parameters, and macros
listings for off-the-shelf applications.
21 CFR Part 11
56
The documentation to be reviewed includes:

technical design-related deliverable
system requirements specification deliverable
system specification deliverables
test plans
integration test results
integration test reports
WHITE BOX TESTING

White box testing (also referred to as structural testing) takes the internal mechanism
(structure) of the software into account, and shows that the software creator followed the
appropriate quality standards. It usually includes inspection (or walk-throughs) of the
program code (code reviews) and development documents.
Code Reviews, Inspections, and Audits
Refer to the previous section.
Unit Testing
The unit test is a white box-oriented test conducted on modules that require a structural
test, and is the lowest level of dynamic testing, the testing of an individual program.
Individual programs can also be called modules or objects. Unit testing involves the
standalone testing of a software unit before it is integrated into the complete software
system or subsystem. Any errors found during testing are removed; the test case that
uncovered the error (and any associated test cases) is then re-executed. Unit testing may
be performed under simulated or emulated conditions. Items that may be checked during
the unit testing include the following:
verification of the functionality as specified in the design document
verifying the results of subjecting the unit to maximum/minimum limits
the reaction to incorrect inputs
exception handling
algorithm checking
adherence to documented standards
Code reviews and unit testing are both essential stages of the SLC. If an organization
follows established and compliant software quality standards, the result will be more
robust programs going into integration testing and, finally, a system with low
maintenance. Proper coding and unit testing are basic steps to ensure that the system
being built will work, once it is put together.
Less extensive functional testing might be warranted than when the code reviews,
inspections, audits, and unit testing are not performed or not available for inspection.
57
BLACK BOX TESTING

Black box testing is also known as functional testing, benchmark testing and, in the
pharmaceutical community, as an operational qualification (OQ). It tests the functionality
and correctness of a computer system by running the integrated software, and is
performed for one of two reasons: defect detection, and/or reliability estimation. It should
cover all functions of the application software that the end-user will use, and will verify
the completeness and accuracy of system requirements used to define the test cases.
A black box test involves providing an input to the system and then examining the
output. The structural elements of the application software are not examined during the
execution of a black box test; these elements are examined during a white box test. An
OQ is a black box (result oriented) system test, designed to determine the degree of
accuracy with which the system responds to a stimulus. The following sections describe
some test cases that an OQ may contain.
Transaction Flow
Transaction flow testing focuses on how transactions are routed within a system, and how
queues affect the systems performance. Transactions are measured on the accurate entry,
storage, and retrieval of data for various user-controlled operations. Some typical
transactions that can be tested include: acknowledgments; receipts and negative
acknowledgments; transactions for operational diagnostics; initialization or reset
transactions for all external interfaces; transactions utilized during system recovery;
transactions used to measure system performance; transactions used to test system
security; and recovery transactions from an external system.
It is necessary to identify how the transaction enters the system, how it leaves the
system, how it merges (and with whom), how it is absorbed, how it splits, and so on.
Queues, as with any resource in a computer system, have a maximum capacity.
Exceeding the capacity of a queue can crash a computer system. In addition to the
capacity, the testing of a queue includes the order in which transactions are taken off a
queue for processing.
Another aspect of transaction flow testing is synchronization. This type of test covers
the merging of two or more transactions.
Domain Testing
Domain testing (also known as boundary testing), is used to verify numerical processing
software. Domain test values are set at the edges of the valid input ranges (i.e., slightly
above and slightly below), and include the biggest, smallest, soonest, shortest, fastest
members of a class of test cases. It should be noted that the presence of incorrect
inequalities usually causes failures only at the code of algorithm boundaries. The
objective of domain testing is to avoid making the computer system crash or
involuntarily stop.
21 CFR Part 11
58
State Transitions
Interactive programs move from one visible state to another. If something changes the
range of available choices or alters the program display on the screen, the programs state
has been changed. A menu system is an example. When an option is selected, the
program changes state, and displays a new menu. Each option and pathway in each menu
must be tested. State transitions are intended to identify any inconsistencies between the
observed functionality and the functionality presented in the system users guide.
Function Equivalence
This test evaluates the functions embedded in the program, usually performed during the
unit testing phase, but for critical functions it can be executed again during the
operational qualification. The function in the program being tested is called the test
function. The calculation to be compared with the test results can be either calculated by
hand or by executing a validated test program.
Load Testing
Load testing is also known as stress testing and includes test cases that cover boundary
loading conditions. Stress tests are designed to challenge programs with exceptional
situations. This includes not only the volume of data a system may process, but also the
rate at which it is processed. A combination of limits is also recommended.
Stress testing executes a system in a manner that demands resources in abnormal
quantities, frequencies, or volumes. Some test cases that can be included are: exceeding
the average rate of interrupts; increasing the rate of data input; requiring the maximum
amount of memory; performing trashing operations in a virtual memory management
scheme; performing excessive seeking operations for disk-resident data; and utilizing the
maximum number of resources.
Configuration Testing
Configuration testing verifies that the software runs on the specified hardware, using the
specified system software and in its specified configuration. Checks for error or warning
messages should be carried out.
System Security Testing
This test focuses on policies, procedures, and restrictions relating to both physical access
to the system hardware, and to on-line program access.
Regression Testing
The testing of modifications to application software and its associated interfaces must be
subject to the same level of testing as performed during software development. The
59
objective is to not only prove the correctness of the modification, but also to uncover any
unforeseen effects in other areas of the computer system.
During regression testing, the previous set of tests that were performed successfully
are executed again. After correcting an error and installing the revised software, the test
cases that caused the error are executed again or the conditions are simulated in order to
try and re-create the circumstances that uncovered the error. In some cases a fix will
have an effect on something else. It is recommended that time is taken to understand the
relationship between the fix and other areas that have already been tested. If necessary,
execute the test cases that are related to the affected item again.
Storage Use Testing
For processes in which a high volume of critical data are gathered, it is necessary to
determine the adequacy and reliability of the storage system. The adequacy of the storage
system is based on the volume of data to be stored and the user requirements.
The FDA publication CPG 7132a.07 provides guidance on the reliability testing
associated with computer storage. This guideline requires that computer I/Os are checked
for accuracy. This testing is performed in two phases. The first phase consists of
qualifying the inputs to the storage system. The second phase requires the
implementation of an ongoing monitoring program to periodically verify the I/Os.
Storage use testing is also associated with the Data Archiving Standard Operating
Procedure. Primary storage is expensive, which is why data is normally transferred to a
less expensive secondary storage media at various times during the execution of the
application software. Data transfer procedures should be tested on the guidance provided
in FDA publication CPG 7132a.07.
OTHER TESTING TYPES

Program Build Testing
Program build testing, which is also known as integration testing, is the incremental
testing performed on units of code (modules), integrated units of code, and the program
as a whole in order to uncover any errors associated with either program or hardware
interfaces. Unit tested modules are built into a program structure that has been developed
to meet the design requirements. The characteristics that are evaluated during the
program build testing are:
interface integrity
functional validity
information content
The software interfaces can be internal and/or external to the application program.
Internal interfaces are those occurring between the application modules or the hardware
components. External interfaces are interfaces to other systems. Guidance on software
integration testing is provided in FDA publication CPG 7132a.07. According to this
21 CFR Part 11
60
guideline, the data input and output checks for automated systems are necessary to ensure
the accuracy on the software interfaces.
Chapter 10
Qualifications
INTRODUCTION
Qualification is the process of demonstrating whether a computer system and its
associated controlled process/operation, procedural controls, and documentation are
capable of fulfilling the specified requirements. Sidebar 1011 depicts the context of the
qualification activities and the computer systems validation process. The physical
components of the systems are reviewed and its specific elements, subunits, and
parameters are documented. A qualification also serves to verify and document the
acceptability of the design, implementation, integration, and installation of the computer
system components. The execution of the qualification is the mechanism by which
documented evidence is created that verifies that the computer system performs
according to its predetermined specification.
1
Courtesy of EduQuest, Inc.
Qualifications
63
Sidebar 101.
Qualification practices apply to all components of a computer system and its associated
controlled process/operation, procedural controls, and documentation. Any of these
components can be purchased or developed by the company or by a computer technology
supplier/contract developer.
The extent of computer technology qualifications depends on a number of factors,
including the criticality and the complexity of the system, whether the qualification is to
be prospective or retrospective, and whether innovative elements are incorporated. Any
qualification activity must be conducted before the technology is made available in the
operating environment or, in case of manufacturing systems, available for process/
product validation. The qualification must be conducted in operating environment to
ensure that the production component, data, and system is not corrupted or compromised.
The qualification of computer hardware, software and the verification of associated
documentation are performed according to the applicable sections in Chapter 7.
21 CFR Part 11
64
This chapter describes the general practices for the qualification of computer systems
(system software, application software and hardware) that perform regulated operations.
The following areas are discussed:
hardware installation qualification
software installation qualification
system operational qualification
system performance qualification
operating systems and other system software qualification
standard instruments, microcontrollers, and smart instrumentation qualification
standard software packages or off the shelf software qualification
configurable software qualification
custom-built systems qualification
The information contained in this chapter is written by enumerating the key practices that
relate to specific qualifications.
HARDWARE INSTALLATION QUALIFICATION2
Objective
Hardware installation qualification of the computer technology provides documented
evidence of a high degree of assurance that the hardware has been installed according to
an appropriate and approved technical design specification, and the hardware suppliers
installation manual(s). Appropriate personnel must conduct the installation qualification
of the computer systems hardware.
The term computerized hardware (hereafter referred to as hardware) refers to
physical elements including input devices, output devices, network components, signal
converters and computers (e.g., controllers, microprocessors, smart instruments).
Practices
The following practices are applicable to all hardware components considered part of a
computer technology performing regulated operations:
Installation of hardware components that are part of a computer system are qualified
according to procedural controls.
Requirements for verifying and recording the results of the installation of computer
hardware must be contained in procedural controls.
Hardware installation test cases may be developed with the collaboration of the
application support personnel, the application expert, or system user.
2
O.Lpez, Regulatory Systems Hardware Qualification, Sue Horwood Publishing Limited, ISBN
1-904282-05-9 (http://www.suehorwoodpubltd.com/).
Qualifications
65
Equipment separate from the computer system but considered part of the computer
system must be determined (if calibration and qualification is required) prior to, or as part
of, the computer system installation qualification; this situation may arise in certain
laboratory or shop floor situations.
If calibration of ancillary equipment or tools used for installation is required, then the
measurement methods, traceability of the calibration standard to international, national,
or approved standards, and equipment specifications must be provided for acceptance.
If a validation plan is available, the hardware installation qualification should be
addressed in this plan. If there is no validation plan, the hardware installation plan may be
addressed either in a standalone document or as part of the installation qualification
protocol. The hardware installation qualification plan identifies the activities to be
completed and the responsible parties involved.
Hardware installation activities and documentation must be clearly communicated to
the hardware/software supplier(s) before reaching any contractual agreements.
Before the execution of the hardware IQ, checklists, tables, and/or forms must be
created to collect the test result data. These are approved as part of the IQ and completed
during the qualification. The IQ protocol covers all installation activities and may
include, but not be limited to, the following:
bill of materials
environmental measurements
design specifications and acceptance criterion
verification of the as installed diagrams
installation verifications
date and time of execution
personnel performing the qualification/verifications
signatures of installation personnel
acceptances signatures of appropriate personnel
The inspection of all hardware must be indicated and documented, including the
equipment received, the personnel performing the activity, and the date of the activity. A
complete list of the hardware to be qualified must be created, including the hardware
item, description, manufacturer or supplier, manufacturers model number, and serial
number.
Hardware configuration diagrams must be retained as part of the qualification
documentation. This is especially important in certain laboratory and shop floor situations
where the computer hardware is essential to the operation of the equipment.
A checkout of critical input/output and electrical connections must be documented,
including (as applicable):
wiring connections and I/Os
radio frequency transmission devices
shielding and wire type specifications
grounding
critical signal levels
verification of as-built wiring diagrams
Electrical specifications must be documented and verified, including:
21 CFR Part 11
66
voltage
amperage
phase requirements
power conditioning requirements
backup power-supply requirements
For computer workstations, the workstation configuration must be verified using the
technical design specification. The result of the verification must be documented using
the data collection forms located in the approved protocol.
Environmental requirements and specifications must be documented and assurance
provided that the environmental parameters are within the range specified by the
manufacturer.
Critical environmental requirements may exist in the following areas:
temperature (e.g., air, fluid, hydraulic, etc.)
humidity
electrical enclosure ratings
airflow
particulate control
pressure
cleaning requirements
spurious radio frequency interference
electromagnetic interference
Verification must be provided for the approach utilized, to mitigate any items identified
in the risk analysis that is related to the hardware.
The physical security arrangements for the hardware must be verified. Critical
hardware must be placed in a physically secure area to prevent any unauthorized physical
access.
It must be verified that the hardware installation adheres to the applicable safety
standards and specifications in the areas of:
fire safety
OSHA regulations
electrocution hazards
ergonomic requirements
Verify that the equipment has been tagged or identified according to the applicable
policies and procedures.
Verify spare parts lists (if applicable).
Verify the hardware specifications for routine and preventive maintenance.
Execute any supplied diagnostics to ensure proper installation.
The installation qualification will include, but will not be limited to, the following
technologies/documentation:
infrastructure
storage devices
colour graphics and displays
printers
Qualifications
67
hard disk and floppy systems

I/O termination units
signal conditioners and transducers
sensors and control elements
field devices
electrical services (e.g., power sources, UPS, grounding, shielding)
as built drawings
supplier SLC documentation supplied under the contract
customer documentation (e.g., user requirements, validation plan, supplier audit, and
supplier selection, as applicable)
Once the hardware installation qualification protocol has been completed, the test results,
data and documentation are formally evaluated. The written evaluation should be
presented clearly and in a manner that can be readily understood. The structure of the
report can parallel the structure of the associated protocol. The report should also address
any nonconformances encountered during the hardware installation qualification, and
their resolution. The hardware installation qualification report summarizes the results of
the verification and testing of all hardware technologies that are part of the system.
The Hardware Installation Qualification Report can either be a standalone document
or may be incorporated into the overall system qualifications summary report.
The approved Hardware Installation Qualification Report should be submitted for
retention to the appropriate record retention organization.
Networks are generally extensions of distributed processing. They may consist of
connections between complete computer systems that are geographically distant, or they
may consist of computer systems on a LAN in the same facility.
In Guide to Inspection of Computerized Systems in Drug Processing, the FDA raises
the following issues concerning networks:
What outputs, such as batch production records, are sent to other parts of the network?
What kinds of input (e.g., instructions, software programs) are received?
What are the identities and locations of establishments that interact with the firm?
What is the extent and nature of the monitoring and controlling activities exercised by
remote on-net establishments?
What security measures are used to prevent unauthorized entry into the network, and
what are the possibilities for unwarranted process alterations, or the obliteration of
process controls and records
For network implementation purposes, the first decision is whether the network is
considered an open or a closed system. The open/closed system issue will drive the
technological and procedural controls that need to be considered during the system
implementation, and that need to be addressed as part of the system risk analysis.
21 CFR Part 11
68
SOFTWARE INSTALLATION QUALIFICATION

Objective
The installation qualification performed on the applications and the system software
provides documented evidence to show, with a high degree of assurance, that the
software has been installed according to an appropriate and approved technical design
specification, and to the suppliers installation manual(s). Appropriate personnel must
conduct the installation qualification of the application and system software.
Table 101. Software IQ Deliverables and

Acceptance Criteria.
Deliverable
Acceptance Criteria
Criticality analysis
report
System
requirements
System
specification
Technical design
specification
Traceability
analysis
verified and approved document

verified and approved as-built document
verified and approved as-built document that traces

system requirements to specific section/clauses in the system
specification, technical design specification, development
testing, and acceptance testing documentation
Assessment(s) and verification that the SLC documentation conforms to the users
progress audit
requirements and the suppliers quality procedures
reports
Test plan
verified to be in conformance with the as built system
Training plan
verified and approved for execution
System
installed as designed
environment and
related utilities test
results
System physical
security
User guides and
verified and approved, as-built
manuals
Technical guide
Installation and
maintenance
instructions
Completed
installed as designed, and successful startup and shutdown of the
hardware
computer system
installation
Qualifications
System software
installation
Program source
code (if applicable)
System
configuration
Application
installation and
inventory
Initial system data
(data conversion, if
applicable)
System
management
procedural controls
Maintenance and
log book
Service level
agreements
69
installed as designed, and successful startup and

shutdown of the software
conform copy of as delivered, and successful startup
and shutdown of the software
verified and approved as-built
verified against the contract and approved
Practices
The following practices are applicable to all software components considered part of all
computer systems performing regulated operations:
The installation of software components that are part of a computer system are
qualified according to procedural controls.
The requirements for verifying and recording the results of the installation of
computer application and system software must be contained in procedural controls.
The software installation test cases may be developed with the collaboration of the
application support personnel, the application expert, or a system user.
If a validation plan is available, the software installation qualification should be
addressed in this plan. If there is no validation plan, the software installation plan may be
addressed in either a standalone document or as part of the installation qualification
protocol. The software installation qualification plan identifies the activities to be
completed and the responsible parties.
Table 101 lists the deliverables to be gathered and verified during the installation of
system and application software. The results of this activity must be recorded in the IQ
protocol or similar document.
The list below also includes some brief acceptance criteria, but this should not
necessarily be considered to be a complete description.
The software installation qualification can be executed in parallel with, or after, the
hardware installation qualification.
If applicable, software virus checking must occur before the installation qualification
of the software.
Before the execution of the software IQ, checklists, tables, and/or forms must be
created to collect the test results data. These are approved as part of the IQ protocol and
21 CFR Part 11
70
completed during the qualification. The IQ protocol covers all of the installation activities
and may include, but should not be limited to, the following:
personnel performing the activities
computer hardware and software installation activities
recording of software and hardware configuration parameters
verification of installation requirements, including database structure, files, tables, and
dictionaries, as applicable
provision of appropriate utilities to demonstrate the correct installation/connectivity
verification of environmental and power utilities
complete list of the software to be qualified is created including:
name and version of the product
brief description of the software (e.g., operating system, document management
application package, custom application, etc.)
suppliers names
license numbers or other identifying information (if applicable)
Prior to software installation, a back-out or roll-back plan and procedure must be
documented. If the planned installation fails, the back-out plan is implemented and the
restoration activities and results are documented. The back-out plan includes
information on performing a backup of the system, and/or making use of an existing
backup of the system, which has been verified prior to software installation.
Upon successful installation of the software, a printout of the resulting directory
listings may be obtained for verification and documentation purposes. This printout may
be enclosed as part of the executed software IQ protocol.
If possible, a printout of the installation audit or system log file should be captured
at the time of software installation and retained as documented evidence. In some cases,
the supplier may provide internal diagnostic software that verifies the successful
installation/connection of the computer technologies. A printout of any such diagnostic
test results or reports should be retained as evidence. If it is not possible to capture this
evidence, then the expected results of a successful installation must be indicated in the IQ
protocol, and the actual observed results documented.
All installation qualification activities and resulting data must be verified by a second
(independent) person. Evidence of verification activities may be confirmed by signing of
the corresponding data collection form; verification activities must include the date on
which the activity was performed. The reviewers signature indicates that the test was
completed as expected, that the acceptance criteria were met, and that the appropriate
documentation/ evidence (if applicable) was collected.
If nonconformances or deviations to the project plan or IQ protocol are encountered
during the software IQ testing, these must be documented, analyzed, resolved, reviewed,
and approved. The resolution process must indicate what additional actions must be taken
to provide a conforming product (e.g., return the program to development for error
analysis and correction, and re-execute the test script after correction; the
nonconformance was due to an inaccuracy in the test script, review and update the test
script, etc.). After the successful resolution of the nonconformance has been obtained, the
original test, the nonconformance information, and the retest must all be retained,
approved, and reviewed by the appropriate personal.
Qualifications
71
Once the software installation qualification protocol has been completed, the test
results, data, and documentation are formally evaluated. The written evaluation should be
any nonconformances encountered during the software installation qualification and the
resolution. The software installation qualification report summarizes the results of the
verification and testing of all technologies that are part of the system.
The Software Installation Qualification Report can either be a standalone document or
may be incorporated into the system qualification summary report.
The approved Software Installation Qualification Report should be submitted for
retention by the appropriate records retention organization.
SYSTEM OPERATIONAL QUALIFICATION

Objective
The system operational qualification performed on the computing technology provides
documented evidence of a high degree of assurance that the hardware and software
components perform as required by the system specification. This includes the
verification that each unit or subsystem of the system operates as intended throughout all
anticipated operating ranges. Appropriate personnel must conduct the system operational
qualification.
Practices
The following practices are applicable to all system components considered part of
computer technologies which perform regulated operations:
The system components that are part of the computer technologies are qualified
according to procedural controls.
The requirements for testing and recording the results of the operation of the computer
technologies must be contained in procedural controls.
The system operational test cases may be developed with the collaboration of the
application support personnel, the application expert, or a system user.
If a validation plan is available, the system operational qualification should be
addressed in this plan. If there is no validation plan, the system operational qualification
plan may be addressed in either a standalone document or as part of the system
qualification protocol. The system operational qualification plan identifies the activities
to be completed and the responsible parties involved.
Before execution of the operational qualification, an installation qualification must be
performed on all relevant computer technologies.
The application support personnel, the application expert, or a system user may
develop the system operational qualification test cases.
Components that require calibration are identified and calibrated by qualified
personnel before, or as part of, the installation qualification.
21 CFR Part 11
72
Before the execution of the system OQ, checklists, tables, and/or forms must be
created to collect the test results data. These are approved as part of the OQ protocol and
completed during the qualification. The OQ protocol covers all of the system operational
activities and may include, but should not be limited to, the following:
personnel performing the activities
testing of the software and hardware configurable parameters
the use of any data migration tools. These should be written in accordance with an
established quality management system. The data migration process should be verified
and the verification performed before final data loading for release
verification of applicable functions:
verification of the executed computer technologies installation qualification
audit trials
electronic signatures
alarms, alarms actions, events, interlocks, and safety circuit testing
confirmation of data acquisition
algorithm, data handling, and report generation
timers (if applicable)
power failure and recovery
disaster recovery
proper communications between the computer system and its peripherals
restart and recovery functions and procedures
availability of an application for authorized accounts holders
integration with other computer system
As explained in Appendix E, the content of the FAT, SAT, and OQ can be similar. The
system specification forms the basis for these activities.
Taking into consideration the previous statement, for the purpose of project
management the operational qualification may be divided into two (2) main types: the
OQ1 or FAT, and OQ2 or SAT. See Figure 101. Additional information can be found in
Appendix E.
Figure 101. Qualification Timeline.

In very large and complex systems developed by contract developers, it is the practice to
perform an FAT or OQ1. The FAT is performed in the development environment by the
Qualifications
73
supplier/integrator on the suppliers/ integrators premises in order to verify the proper

operation of the application before delivery to the users site. If the FAT results comply
with the system specification, the system is released for shipment and installation on the
users site.
All critical deviations found during the execution of the FAT must be corrected before
system release and shipment.
Regression testing must be executed in order to verify that all critical deviations were
corrected.
In all large and complex systems developed by contract developers, it is normally the
practice to perform an SAT. The SAT is the set of activities comprising the system
installation, start-up, operational testing, and user orientation performed by the software
supplier/integrator in the operational environment. For the purposes of this book, the
operational testing performed as part of the SAT is also known as the OQ2. If the SAT is
conducted according to approved protocols, test results are properly documented, and test
results reviewed and approved by the user and QA, it may not be necessary to repeat the
installation and operational qualifications. This provides cost savings and can streamline
the validation process. Refer to Sidebar 102.
THE OQ (1) and (2) CONCEPT
OQ1 refers to the testing performed in a developmentenvironment, which can be on
the suppliers premise,or within the users company. The
supplier/integratorperforms this testing and it is witnessed by thecompany
personnel, including QA. OQ1 is oftenreferred to as an FAT.
OQ2 refers to the testing performed in an operational environment, usually the
environment where the system will be used when operational, but before official
release. The supplier/integrator performs this testing and it is witnessed by the
company personnel, including QA. OQ2 is often referred to as a SAT.
This concept applies to all technologies of a computer system.
APPLYING THE OQs (1) and (2) CONCEPT
Since OQ2 witnessed testing takes place under operational conditions, there is
no need to give guidance on the control of the testing environment. The IQ will
provide evidence that the installation is appropriate for operational use.
During the execution of OQ1, in the development environment, it should be
ensured that this environment models the specified operating environment. This is
because unrealistic or unconvincing test environments can make tests unsuitable as
evidence of proper operation. Unrealistic tests could be tests that are performed on
old program versions, or where there is no evidence of regression testing for the
program. When consolidating of OQ1 and OQ2 tests, only the tests that have been
identified as suitable should be taken into account.
Sidebar 102.
21 CFR Part 11
74
All system operational qualification activities and resulting data must be verified by a
second (independent) person. Evidence of verification activities is confirmed by signing
of the corresponding data collection forms. The verification activities must include the
date that the activity was performed; the reviewers signature indicates that the test was
completed as expected, that the acceptance criteria were met, and that the appropriate
documentation/evidence (if applicable) was collected.
If nonconformances or deviations to the plan or protocol are encountered during the
system OQ testing, these must be documented, analyzed, resolved, reviewed, and
approved. The resolution process must indicate what additional actions must be taken to
provide a conforming system (e.g., return the program to development for error analysis
and correction, and re-execute after correction; the nonconformance was due to an
inaccuracy in the test script; update the test script; and so on). After the successful
resolution of the nonconformance, the original test performed, the nonconformance
information obtained, and the retest must all be retained, approved, and reviewed by the
appropriate person.
Once the system operational qualification protocol has been completed, the test
any nonconformances encountered during the system operational qualification and their
resolution. The system operational qualification report summarizes the results of the
verification and testing of all technologies that are part of the system.
The System Operational Qualification Report can either be a standalone document or
may be incorporated into the overall system qualification summary report.
The approved System Operational Qualification Report should be submitted for
retention to the appropriate records retention organization.
Upon successful completion of the operational qualification, the computer system is
available for further qualification (PQ).
SYSTEM PERFORMANCE QUALIFICATION

Objective
Computer systems performance qualification performed on the computing technology
provides documented evidence of a high degree of assurance that the computer system
and the operating environment (Figure 21) perform as established by the requirements
specification deliverable.
Appropriate personnel must conduct the system performance qualification.
Practices
The following practices are applicable to the integrated computer system (which include
the controlled process, operating procedures, and supporting documentation) that
performs regulated operations:
Qualifications
75
The integrated computer system is qualified according to procedural controls

The requirements for testing and recording the results of the integration of the computer
system must be contained in procedural controls
The system integration test cases may be developed with the collaboration of the
application support personnel, the application expert, or a system user
If a validation plan is available, the system performance qualification should be
addressed in this plan. If there is no validation plan, the system performance qualification
plan may be addressed in either a standalone document or as part of the system
qualification protocol. The performance qualification plan identifies the activities to be
completed and the responsible parties involved.
The application support personnel, the application expert, or a system user may
develop the system performance qualification test cases.
Before the execution of the system performance qualification, checklists, tables,
and/or forms must be created to collect the test results data. These are approved as part of
the PQ protocol and completed during the qualification.
The specific objectives of the system performance qualification include:
determining the accuracy of the integrated system in the receiving, recording, storing,
and processing of data
determining the accuracy of the integrated system in arriving at an appropriate decision
based upon the data received and the established decision matrix
For a process control system, the system performance qualification may involve
simulation run(s) on the controlled process and its associated equipment using placebo or
process materials. These runs are not the same as the product/process performance
qualification, although they can be used to support this activity. System performance
qualification demonstrates and verifies computer system suitability. Product/process
performance qualification demonstrates and verifies product/process suitability.
For business/corporate computer installations, it may be difficult to separate system
operational qualification and system performance qualification. Procedural controls may
be used to address this issue.
All system performance qualification activities and resulting data must be verified by
a second (independent) person. Evidence of the verification activities are confirmed by
signing the corresponding data collection form. The verification activities must include
the date that the activity was performed. The reviewers signature indicates that the test
was completed as expected, that the acceptance criteria were met, and the appropriate
documentation/evidence (if applicable) was collected.
If nonconformances or deviations to the project plan or protocol are encountered
during the system PQ testing, these must be documented, analyzed, resolved, reviewed,
and approved. The resolution process must indicate what additional actions must be taken
to provide a conforming system (e.g., return the program to development for error
analysis and correction, and reexecute the test script after correction; the nonconformance
was due to an inaccuracy in the test script, update the test script, etc.).
After the successful resolution of the nonconformance, the original test performed, the
nonconformance information obtained, and the retest must all be retained, approved, and
reviewed by the appropriate personal.
21 CFR Part 11
76
Once the system performance qualification protocol has been completed, the test
any nonconformances encountered during the system operational qualification and their
resolution. The system performance qualification report summarizes the results of the
verification and testing of all technologies that are part of the integrated system.
The System Performance Qualification Summary Report can either be a standalone
document or may be incorporated into the overall system qualification summary report.
The approved System Performance Qualification Summary Report should be submitted
for retention to the appropriate records retention organization.
Upon successful completion of the performance qualification, process control systems
are available for process/product performance qualification and business computer
systems are available for operation.
OPERATING SYSTEM AND UTILITY SOFTWARE

INSTALLATION VERIFICATION
Introduction
The practices contained in this section apply to all operating systems and system software
components that are considered part of a computer system.
System software (e.g., operating systems, assemblers, utilities) is defined as:
independent software that supports the running of application software (ISO)
software designed to facilitate the operation and maintenance of a computer system and
its associated programs (IEEE)
Objectives
All computer systems performing regulated operations must be validated. To support
computer systems validation, operating systems and system software components are
installed according to the manufacturers instructions. The installation verification
includes documentation gathered during the verification activities.
Practices
The installation of system software and its associated components is verified according to
normal software installation qualification practices. (Refer to Software Installation
Qualification in this chapter.)
Well-known operating systems do not need functional testing to be performed by the
enduser.
Due to the close relationship between the computer hardware and the system software,
operating system installation verification may be performed during the hardware
installation qualification activities.
Qualifications
77
New versions of system software (e.g., operating systems, assemblers, and utilities)
should be reviewed prior to use and consideration should be given to the impact caused
by any new, amended, or removed system software features on the application(s) that run
on it. This impact assessment could lead to formal regression testing of the application(s),
particularly when a major upgrade of the operating system has occurred. The results of
the regression testing must be documented and approved by the appropriate personnel:
The following project activities described in Chapter 7 are also applicable to system
software and system software components:
Conduct a system (hardware, software and process) risk study and a
criticality/complexity assessment. Usually this study and assessment are performed to
quantify the impact of new versions on the actual configuration
Complete change control forms and approvals, including obtaining the approval of QA
Preparation of the installation qualification protocol, if applicable, which includes
activities associated with the regression testing of the applications impacted by the
installation of system software
Execute the IQ protocol
If applicable, perform regression testing of the application programs (OQ and PQ)
Prepare a summary report
Update the documentation regarding references to system software, system software
components, and their associated versions
The approved summary report and associated validation documentation should be
submitted for retention to the appropriate records retention organization
Part 11 Areas of Interest
The system administrators position requires the ability to access all items in the
operating system. Key areas of concern are the protection of records, the integrity of the
application software files, and the operating system configurable parameters. In addition
to the security of the system, the Part 11 areas to be considered for an operating system
are as follows:
Are the records protected against intentional or accidental modification or deletion?
Is an electronic audit trail function automatically generated for all system administrator
entries?
Is the audit trail completely outside of the control and access of the system
administrator (except for the read-only access of the audit trail file)?
Is it impossible to disable the audit trail function without being detected?
Is the system date and time protected from unauthorized change?
Is the audit trail data and/or metadata protected from accidental and/or intentional
modification or deletion?
Is the distribution of, access to, and use of operating system and maintenance
documentation controlled?
21 CFR Part 11
78
STANDARD INSTRUMENTS, MICROCONTROLLERS, SMART

INSTRUMENTATION VERIFICATION
Introduction
The practices contained in this section apply to all standard instruments, microcontrollers,
and smart instrumentation (e.g., weigh scales, bar code scanners, controllers, vision
systems and EPROMs) considered part of a computer system. This equipment is driven
by programmable firmware.
Objective
computer systems validation, standard instruments, microcontrollers, and smart
instrumentation are functionally tested (black box test) and the installed versions
recorded.
Practices
The installation of standard instruments, microcontrollers, smart instrumentation, and its
associated components is verified according to normal software installation qualification
practices. (Refer to Software Installation Qualification in this chapter.)
Since the users company does not have control over the system product to be
acquired, the users company must adequately understand and specify the requirements
and capabilities to be provided by the system product.
Standard instruments, microncontrollers, and smart instrumentation can be
configurable. The configuration must be recorded using an IQ protocol, including the
version of the hardware, versions of the firmware (if applicable) and associated software.
The introduction of new versions of firmware or software must be performed through
the completion of rigorous change control.
Backups of firmware or software are critical items for operational continuity.
The impact of new versions of firmware and software on the validity of the existing
qualification documentation should be reviewed and appropriate action taken. If
applicable, after the new version is installed, regression testing should be conducted. The
results of the regression test must be documented and approved by the appropriate
personnel.
If a manual system is being replaced by one of these technologies, the two systems
should be run in parallel for a time as a part of the performance qualification.
The operational qualification for standard instruments, microcontrollers, and smart
instrumentation consists of a black box test. This type of test is based on the users firm
application requirement and challenges a programs external influences. It views the
software as a black box concerned with program inputs and its corresponding outputs.
The black box testing must consider not only the expected (normal) inputs, but also
unexpected inputs. Black box testing is discussed in Chapter 9.
Qualifications
79
If data is recorded onto magnetic media (electronic records), then the following will
apply:
Records should be secured by physical or electronic means against wilful or accidental
damage. The records collected should be checked for accessibility, durability, and
accuracy. If changes are proposed to the computer equipment or its programs, the
above mentioned checks should be performed at a frequency appropriate to the storage
medium being used.
Records should be protected by backing up at regular intervals. The backups should be
stored for as long as is necessary to conform to regulatory and company requirements,
and at a separate and secure location.
There should be adequate alternative arrangements made available for systems in case of
a breakdown. The time required to bring the alternative arrangements into use should be
related to their criticality and the possible urgency of the need to use them.
The procedural controls to be followed if the system fails or breaks down should be
defined and qualified. Any failures and remedial actions taken must be recorded.
Procedural controls must be established to record and analyse errors, and to enable
corrective action to be taken.
The following project activities described in Chapter 7 are also applicable to standard
instruments, microcontrollers and smart instrumentation:
Define the scope of the system and its strategic objectives
Perform a preliminary evaluation of the technology
Provide a Process Description including process flow diagrams and process and
instrumentation drawings (P&IDs)
Complete the change control forms and approvals, including obtaining the approval of
QA
Prepare a system requirement specification deliverable
Prepare and approve a Validation Plan
Conduct a System (hardware, software and process) risk study and a
Criticality/Complexity Study and document the results. Any new requirements that
result from the risk study must be added to the system requirements specification
deliverable
Approve the system requirements specification deliverable
Select the supplier
Revisit the risk study
If changes to the hardware and/or software are required, develop and approve the
hardware/ software design, implement the hardware and software using the design
specification, and conduct an integration test
Revisit the risk study to confirm that any risks have been addressed
Prepare validation/qualification protocols
Conduct an IQ
Conduct an OQ/PQ (usually these can be carried out together)
Develop and approve a summary report
If applicable, and as part of the manufacturing systems interface with the standard
instruments, microcontrollers, or smart instrumentation, conduct a Process/Product
21 CFR Part 11
80
Performance Qualification. Note: This activity is typically performed on

manufacturing systems only. Prepare a Product/Process Report
Update the documentation regarding references to application software and their
associated version
The approved summary report(s) and associated validation documentation should be
submitted for retention by the appropriate record retention organization
Standard Instruments, Microcontrollers, and Smart Instruments are usually connected to a
device to record their data. It is uncommon for them to contain functionality that supports
electronic signatures. The majority of these systems are hybrid systems and if this is the
case, the Part 11 areas to be assessed include:
audit trails and metadata management
system security
access code and password maintenance and security
password assignments
operational checks
authority checks
location checks
document controls
records retention and protection
open/closed system issues
On hybrid systems, another area to be evaluated is the procedural controls on the link
between handwritten, signed (paper-based) report(s) and the electronic records associated
with the report. This link to the record(s) must be based on a unique attribute.
The areas of Part 11 that need to be assessed for Standard Instruments,
Microcontrollers, and Smart Instruments when these devices are not connected to a
recording device include:
system security
password assignments
operational checks
authority checks
location checks
document Controls
Qualifications
81
STANDARD SOFTWARE PACKAGES QUALIFICATION

Introduction
Standard Software Package are products that have already been developed and are usable
as is or with modifications.
The practices contained in this section apply to Standard Software Packages (e.g.,
spreadsheets, text processors, database programs, CAD systems, graphics packages, and
programs for technical or scientific functions) that are considered part of a computer
system. By definition, this includes noncustomized application software that has been
purchased and that has not been modified.
When classifying a software application as either a Standard Software package or
Configurable Software package, consideration should be given the amount of application
development needed for the proposed solution. The classification selected will determine
the validation approach to be taken. For example, SAS is a comprehensive data
management system with modules for application development, spreadsheets, graphing,
data entry, data analysis, project control, business process management, and database
query. If the SAS programs are too complex, SAS may be considered to be a
Configurable Software package instead of a Standard Software package.
Another item of concern regarding Standard Software Packages is that the package by
itself may not be Part 11 compliant. An example of this is a package that does not have
Part 11-compliant audit trails functionality. In this case, the Standard Software package
may need to be combined with third-party applications and/or plug-ins, in order to make
it compliant with Part 11.
A reference to source for the testing of standard software packages is ISO/IEC 12119,
Information technologySoftware packagesQuality requirements and testing. It is
applicable to standard software packages and it provides the following:
a set of quality requirements for standard software packages
instructions on how to test standard software package against specified quality
requirements
a quality system for a supplier who is outside of the scope of ISO/IEC 12119
A RELATED PRODUCT FOR ISO/IEC 12119, THE IEEE

STANDARD ADOPTION OF ISO/IEC 12119
Objective
computer systems validation, the applications developed using standard software
packages are functionally tested (black box testing) and the installed versions recorded.
21 CFR Part 11
82
Practices
Standard software packages are called canned or COTS (Commercial Off-The-Shelf)
configurable packages. The qualification of standard software packages is performed
according to normal software installation and operational qualification practices. (Refer
to Software Installation Qualification and System Operational Qualification in this
chapter.)
Since the users company does not have control over the requirements of the standard
software package, the users company must identify the application requirements and
match them to the available standard software packages.
There is no requirement to validate the standard software package. The validation
concentrates on the elements developed to support the application.
The linking and/or configuration of the standard software package modules is used to
develop the final application.
Standard software packages are configurable, and this configuration must be recorded
in the equipment IQ, along with the version.
Change control must be rigorously applied, since changing these applications is often
very easy, and they have limited security.
The impact (criticality) of new versions on the validity of the qualification
documentation must be reviewed and appropriate action taken. If applicable, after a new
version has been installed, the test cases contained in the operational qualification should
be executed again.
If a manual system is being replaced, the two systems should be run in parallel for a
time, as a part of the operational qualification.
The qualification effort should concentrate on the application itself and should
include:
application requirements and functionality
the high-level language, scripts, or macros used to build the application
critical algorithms, formulae and parameters
data integrity, security, accuracy, and reliability
operational procedural controls
User training should emphasize the importance of change control and the need to
maintain the integrity of these systems.
Data shall be secured by physical and/or electronic means against willful or accidental
damage. Stored data must be periodically checked for accessibility, durability, and
accuracy If changes are proposed to the computer equipment or its programs, the checks
mentioned above should be performed at a frequency appropriate to the storage medium
being used.
Data must be protected by backing up at regular intervals. Back-up data must be stored
for as long as is necessary to comply with regulatory and company requirements at a
separate and secure location.
There should be adequate alternative arrangements made available for systems in case
of a breakdown. The time required to bring the alternative arrangements into use should
be related to their criticality and the possible urgency of the need to use them.
Qualifications
83
defined and training provided. Any failures and remedial actions taken must be recorded.
The operational qualification for standard software packages consists of a black box
test. This type of test is based on the users firm application requirements specification
deliverable and challenges the packages external influences.
The following project activities described in Chapter 7 are also applicable to standard
software packages:
Perform a preliminary evaluation of technology
QA
Prepare a system requirement specification deliverable
Conduct a System (hardware, software, and process) risk study and a
result from the risk study must be added to the system requirements specification
deliverable
Select the software package
Prepare and approve the system specification deliverable in full
Prepare a technical design specification deliverable
critical algorithms, parameters, and formulae
data integrity, security, accuracy, and reliability
Perform design reviews, including technical design inspections
Approve the technical design specification deliverable
Prepare validation/qualification protocols and the applicable procedural controls
Build the application
Inspect the developed code. The focus of this inspection is on the critical algorithms and
parameters
Conduct an IQ
Conduct an OQ and include in this process the operation of the application using the
procedural controls
Prepare a summary report
Update the documentation regarding references to application software, associated
versions, and user manuals
The approved summary report(s) and associated validation documentation should be
submitted for retention to the appropriate record retention organization.
21 CFR Part 11
84

The output from standard software packages is typically recorded to a magnetic media.
Part 11 areas to be assessed include:
audit trails and metadata management
system security
passwords assignments
operational checks
authority checks
location checks
document controls
records retention and protection
If electronic signatures are implemented, the following additional areas must be assessed:
electronic signature security
electronic signature without biometric/behavioral detection
electronic signature with biometric/behavioral detection
signature manifestation
signature purpose
signature binding
certification to FDA
CONFIGURABLE SOFTWARE QUALIFICATION

Introduction
The practices contained in this section apply to Configurable Software (e.g., SCADA
systems, Laboratory Information Management Systems, Manufacturing Resource
Planning Systems, Document Management Systems, Building Automation/Management
Systems, Electronic Batch Recording Systems, Electronic Document Management
Systems, Manufacturing Execution Systems, etc.) that are considered part of a computer
system. The level complexity of the individual programs that are used to form the
released application are a critical factor to consider for this type of software package.
Objective
computer systems validation, the quality of the software embedded in the standard
product must be established. The configurable elements of the standard product must be
reviewed and tested.
Qualifications
85
Practices
Configurable Software packages are also called custom configurable packages.
The software system and hardware platform must be well known and mature before
being considered to be Configurable Software; otherwise the software validation
activities for Custom-Built systems must be used.
A typical feature of configurable software packages is that they permit users to
develop their own applications by configuring/amending/scripting predefined software
modules and by developing new application software modules using the standard (core)
product. Each application (of the standard product) is therefore specific to the users
process and maintenance becomes a key issue, particularly when new versions of the
standard product are released.
Since the users company does not have control over the requirements of the standard
product, the users company must identify the application requirements and match these
to the standard product that are available.
There is no requirement to validate the standard product but the quality embedded in
the standard product must be evaluated as follows:
Conduct a quality audit of the supplier of the standard product to review the level of
quality and structural testing built into the standard product
If the end-user cannot directly review the level of quality embedded in a standard
product, extensive functional testing must be performed in order to assess the quality
of the standard product
The review activities for the standard product should concentrate on the following:
Conduct a feasibility study
Generate a process description which includes process flow diagrams/data flow
diagrams and, if applicable, PI&Ds
Complete the change control forms and approvals, including approvals of QA
Develop system requirements specification deliverable
Conduct a system (hardware, software, and process) risk study and a
result from the risk study must be added to the requirements specification deliverable
Conduct a quality audit of the software supplier (main candidates) who supply the
standard configurable software product that covers:
an audit to establish assurance the suitability of the quality system used to develop
the software (standard product)
review of critical algorithms, parameters, formulae
review of the data flow diagrams and resulting products for data integrity, security,
accuracy, and reliability
review of the Part 11 elements contained in the standard product, such as logical
security
writing and approval of an audit report
21 CFR Part 11
86
selection of the supplier of the standard product

Following the qualification activities used for a custom-built system performs the
qualification of the configurable elements of the standard product. (Refer to Custom-Built
Systems in this chapter.)
The qualification of the configurable elements of the standard product is performed
according to normal software installation, operational and performance qualification
practices. (Refer to Software Installation Qualification, System Operational Qualification
and System Performance Qualification in this chapter.)
Change Control Must Be Rigorously Applied
The impact (criticality) of new versions on the standard product must be reviewed and
appropriate action taken. After the new version has been installed, the new functionality
must be tested and regression testing performed as part of the qualification activities.
User training should emphasize the importance of change control and the procedural
controls to support the system.
Data should be secured by physical and/or electronic means against willful or
accidental damage. Stored data must be checked for accessibility, durability, and
accuracy. If changes are proposed to the computer equipment or its programs, the checks
being used.
for as long as is necessary to comply with regulatory and company requirements at a
defined and qualified. Any failures and remedial actions taken must be recorded.
This item is the same as for Standard Software Packages.
CUSTOM-BUILT SYSTEMS QUALIFICATION

Introduction
Custom-built systems produced for a customer specifically to meet a defined set of user
requirements. The practices contained in this section apply to custom-built applications
Qualifications
87
(e.g., PLC-based systems, the configurable elements within configurable software

applications, any custom-built software applications, and any custom-built interfaces to
other computer systems or databases) that are considered part of a computer system.
Objective
computer systems validation, the custom-built system must be validated. The SLC must
incorporate a formal development methodology with multiple inspections of the
suppliers development and associate testing activities and qualification testing.
Practices
For these systems, the full life cycle must be followed for all elements of the system
(hardware and software).
An audit of the application developer (internal or external) is required to examine their
existing and established quality systems.
A validation plan must be developed, based on the results of the quality audit and on
the complexity of the proposed custom-built system, which describes the development
methodology, inspection, and qualification activities.
The qualification of the custom-built application must be performed according to
normal software installation, operational and performance qualification practices. (Refer
to Software Installation Qualification, System Operational Qualification, and System
Performance Qualification in this chapter.) The following project activities depicted in
Chapter 7 are applicable to custom-built systems:
Conduct a feasibility study
Process description/operation, which include data flow diagrams and, if applicable,
P&ID
QA
Prepare the system requirements specification deliverable
Conduct a System (hardware, software, and process) risk study and
result from the risk study must be added to the requirements specification deliverable
To conduct a quality audit of the supplier of the custom-built product:
Perform an audit to establish the suitability of the quality system used to develop the
software (custom-built system)
Review critical algorithms, parameters, and formulae
Review the data flow diagrams and resulting product for data integrity, security,
accuracy, and reliability
21 CFR Part 11
88
Review the Part 11 elements contained in the custom-built system, such as logical
security
Write and approve the audit report
Select the developer of the custom-built system
Prepare the system specification deliverable
After a technical review, approve the system specification deliverable
Prepare the technical design specification deliverable
Perform a technical review on the design specification
Approve the technical design specification deliverable
Build the custom-built application
Conduct a physical and functional audit
Start the development of the qualification protocols
Conduct a Software Inspection
Conduct Software Module Testing and Integration Testing
Conduct an FAT and SAT
Conduct an IQ
Conduct an OQ
Conduct a PQ
Develop and approve the qualification reports and/or Validation Report
Conduct a Process/Product Performance Qualification. Note: This activity is typically
performed on manufacturing systems only
Prepare a Product/Process Report
Update the documentation regarding references to application software and their
associated versions
The approved summary report and associated validation documentation should be
submitted for retention at the appropriate record retention center.
Change control must be rigorously applied.
The impact (criticality) of new versions on the validity of the existing qualification
documentation must be reviewed and appropriate action taken. After the new version has
been installed, regression testing may be required.
User training should emphasize the importance of change control and the procedural
controls to support the system.
Data should be secured by physical and/or electronic means against willful or
accidental damage. Stored data must be checked for accessibility, durability, and
accuracy. If changes are proposed to the computer equipment or its programs, the checks
being used.
for as long as necessary to comply with regulatory and company requirements at a
Qualifications
89
defined, and specified. Any failures and remedial actions must be recorded and evaluated.
corrective action be taken.
This item is the same as for Standard Software Packages.
Chapter 11
SLC Documentation
REGULATORY GUIDANCE
The deliverables that are important in order to demonstrate that software has been
validated include1:
written design specification2 deliverables describing what the software is intended to do
and how it intends to do it
written test plans based on the design specification deliverables, which include both
structural and functional software analysis
test results3 and an evaluation4 of how these results demonstrate that the predetermined
design specification deliverables have been met
SLC DOCUMENTATION
Computer system documentation means records that relate to system operation and
maintenance, from high-level design documents to end-user manuals. Sidebar 111
illustrates some typical SLC documentation. In the software-engineering context,
computer systems documentation is regarded as software. All regulatory provisions
applicable to software are also applicable to its documentation.
Computer system documents are generated during the project. These documents may be
either printed material or electronic records, such as computer files, storage media, or
SLC documentation
91
film. The more documents that are written, the higher the cost of managing them due to
the increasing difficulty of keeping the documents consistent with the computer system.
System documents must be available if needed for review. Obsolete information must
be archived or destroyed in accordance with a written record retention plan. System
documents can be grouped as depicted in Figure 111.
1 FDA, Guidance for the Industry: Computerized Systems Used In Clinical Trials, April 1999.
2 In this context design specification refers to Table 111 for sample documentation.
3 In this context test results refers to execute qualification protocol(s).
4 In this context evaluation refers to qualification summary report(s).
Group
Sample SLC Documentation
Design and testing

specifications
Functional requirements, hardware and software

design, test plans, test specifications/ scripts,
risk analysis
Qualification records IQ, OQ, and PQ
Operating documents Written procedures, operating manuals,
reference manual
Operating records
Processing log print-out, critical alarm print-out
Maintenance records Program or equipment modification requests, system
and documents
status reports, repair orders, preventive maintenance,
maintenance manuals, configuration
management records
Performance records Performance standard specifications, performance
ogs, performance summary reports
Other records
Service agreements, vendor qualification records,
vendors audit reports
Figure 111. System Documents Grouping.

In addition to the above classifications,
Validation plan
Risk analysis
System requirements specification deliverable(s)
System specification deliverable(s)
Technical design specification deliverable(s)
Source code/configurable code/scripts
Quality audit and technical review reports
Testing plans and reports
unit
integration
functional
FAT and SAT plans and results
Qualification plans
Validation/Qualification results and summaries
Traceability analysis
21 CFR Part 11
92
User manuals
Sidebar
111.
Documentation.
Typical
SLC
system documentation can be classified by the source of the system documents, i.e., the
hardware/software supplier and the customer.
Supplier/integrator items:
development specifications (e.g., detailed hardware and software requirements, overall
software functionality, detailed test specifications)
operational specifications and requirements (e.g., standard software packages and
configurable software documentation)
instruction manuals, training manuals
diagrams, flow charts, and descriptions that define the software modules and their
interactions
test databases
Customer items:
For minimum requirements: refer to Regulatory guidance in this chapter.
Refer to Sidebar 111 for a list of typical system documentation.
All systems-related documentation must be rigorously controlled to ensure that it is
consistent with the system in operation, available for inspection by a regulatory agency in
a timely fashion, and to ensure that the system can be properly operated and maintained.
The quality attributes, applicable to any type of document are that it is: unambiguous,
appropriate, comprehensive, available, accessible, followed, and periodically reviewed.
The content of the system documents must be consistent with the type of software
developed, and the criticality and complexity of the system. In addition, documentation
must satisfy five basic elements; that they are: written, appropriate, clear, verified, and
approved.
Computer system documents, which support system validation, are submitted for
retention to the appropriate document control area and retained for a period at least as
long as specified by the predicated regulations applicable to the system.
Computer system documents contain the following elements.
A title page that contains the following:
a title that clearly indicates the nature of the document
if the document is related to a specific software system, the title also references the
software system and the version of the software system
a revision level number for the document
a date indicating the creation date of the final draft of the document (before the approval
date)
a unique document identifier
clearly designated area(s) for review signatures including a statement that clearly
indicates the significance of the signing (e.g., This document has been reviewed by
), an area for the signature(s), the typed name(s) of the signatories, and the date of
each signing
SLC documentation
93
clearly designated area(s) for approval signatures including a statement that clearly
indicates the significance of the signing (e.g., This document has been approved for
execution by or This validation report has been approved as complete by), an
area for the signature(s), the typed name(s) of the signatories, and the date of each
signing
Each page of the document contains the following:
a header containing an abbreviated indication of the title of the document and the
document identifier number
a footer that contains the revision number of the document and the issue date of the
document (which must correlate with the title page)
a page number or other unique page indicator, which also indicates the total number of
pages(e.g., page x of y)
test scripts and test results that are part of a formal inspections and testing activity.
These items should contain appropriate information that clearly identifies them. This
information may include: an identification number traceable to a requirement
identification number, the revision number of the document, the test execution or
review date, and the name and signature of the person who conducted the test or
review
The completion of computer systems documentation should follow a set of predefined
rules as follows:
Always use water-insoluble ink of suitable darkness to ensure copying will accurately
transcribe the numbers
Pencils, white-out, or erasable pen must never be used
Corrections should be made with a single line through the error followed by the initials
and date of the person making the correction
Blanks, lines, or spaces for recording data or comments that were not used when the
document was completed should be crossed through or otherwise indicated as not
applicable
Another, independent, person should verify calculations, algorithms, SLC documentation,
source code, test script execution, test results, validation plans, and qualification
protocols, and so on, as applicable.
Chapter 12
Relevant Procedural Controls
The integrity of the information managed by a computer system is protected by
procedural controls, rather than the technology used to apply the controls. Procedural
controls comprise any measures taken to provide appropriate instructions for each aspect
of system development, operations, calibration, and validation. For computer systems,
procedural controls address all aspects of software engineering, software quality
assurance, and operation. In a regulated environment, these controls are fundamental to
the operation of the computer system.
Procedures covering computer technologies may include:
the operation and maintenance of the technology
the authentication of users, access controls
management of records and audit trails
physical security disaster recovery
protection of remote access points
protection of external electronic communications
personnel responsibilities, and the ongoing monitoring of computer systems
One of the most important procedures to be developed and followed is the administration
and retention of electronic records. Part 11 requires the retention of electronic records in
electronic form. For regulated systems in which electronic signatures are not
implemented (hybrid systems) the electronic record requirements (Sub-Part B) in Part 11
are applicable, and the electronic records are maintained and retained in electronic form
for the period established by the predicate rule.
The procedural controls applicable to computer systems must have QA controls
equivalent to other regulated operations. If personnel do not have complete instructions
or do not understand what they need to do, the outcome can be unexpected results or
operation failures.
The quality attributes applicable to all procedural controls are that they are
appropriate, clear, and consistent with related procedures, adhere to regulatory, industry,
and company standards, and are approved, available, accessible, followed, controlled, and
periodically reviewed. The SLC process must be consistent with the applicable computer
systems validation procedural controls.
Appendix F contains some sample procedures that are mapped to Part 11.
Written procedural controls for computer systems should be available for the
following activities:
the development and maintenance of computer systems
the operation of computer systems
managing the access to the computer system and management of the records
21 CFR Part 11
96
retention and storage of data

prevention of unauthorized access to the data
loss management procedures to deauthorize lost, stolen, missing, or otherwise
potentially compromised tokens, cards, and other devices
system restart and the recover of data
the management of electronic signatures
All procedural controls must be reviewed periodically in order to ensure that they remain
accurate and effective. Where appropriate, a number of topics may be combined into a
single written procedure. Wherever possible, procedures should be established at a level
allowing them to be utilized across a number of implementation and maintenance projects
and systems. When new procedures are needed, existing written procedures are reviewed
to determine if they can be upgraded to include new requirements or processes.
Redundant procedures must be avoided.
The project team and validation coordinator should determine which procedures are
appropriate for a particular system development, validation, and/or maintenance activity.
Chapter 13
Change Management
INTRODUCTION
Adherence to change management practices for computer technologies provides a process
by which a change to a computer system must be proposed, evaluated, approved or
rejected, scheduled, tracked, and audited.
Regulatory guidance
Systems should be in place to control changes and evaluate the extent of
revalidation that the changes would necessitate. The extent of revalidation will
depend upon the changes nature, scope, and potential impact on a validated
system and established operating conditions. Changes that cause the system to
operate outside of previously validated operating limits would be particularly
significant.
FDA, Guidance for the Industry: Computerized Systems Used In Clinical Trials,
April 1999.
As required by the regulations, a procedural control for the implementations of changes
must be defined and documented. A change may be requested due to:
perfective maintenance or to correct the system due to new requirements and/or
functions
adaptive maintenance or to correct the system due to a new environment that could
include new hardware, new sensors or controlled devices, new operating systems,
and/or new regulations
corrective maintenance or to correct the system due to the detection of errors in the
design, logic, or programming of the system
preventive maintenance or to correct the system in order to improve future
maintainability or reliability, or to provide a better basis for future enhancements
Change requests can result from both planned and unplanned events. Planned events can
be scheduled, and can be the result of a planned expansion in business requirements.
Emergency or unplanned events can result from software or hardware problems that
prevent the application from performing its defined tasks.
Changes can be further classified as major or minor. The SLC activities to be
performed are determined by the criticality and complexity of the system and/or the
change. Responsible persons from QA, manufacturing, engineering, R&D, and/or IT
perform the evaluation of the changes. A traceability analysis can help to determine the
impact of the change(s).
21 CFR Part 11
98
The replacement of hardware requires a formal IQ and, if applicable, regression

testing to verify the correct integration of the hardware with the system and application
software. When the hardware is replaced by an equivalent item (e.g., same model,
revision, and manufacturer) it is necessary to record the details of the replacement part.
Emergencies can arise that require an immediate or emergency change in order to
prevent the disruption of business processes.
Changes and modifications to application software can introduce as many
problems/errors as they solve unless they are rigorously controlled. Site changes and
modifications to application software are particularly difficult to manage due to the lack
of a formal, quality controlled, development environment.
CHANGE MANAGEMENT PROCESS

The following process is applicable for changes to any component of computer
technology that performs functions on regulated products, including hardware, peripheral
devices, system software and utilities, application software, documentation, and
communications hardware and software. These components can be either purchased from
an external supplier or developed in house by the company or its designee.
Change management procedural control must be in place when the validated system is
released for use. These controls must address the identification and specification of the
change.
Initiation is the process by which a change is requested. This process includes:
documenting the change required; providing a rationale for the change and its anticipated
impact on other areas of the system hardware and/or software; and forwarding the change
request to the appropriate personnel (e.g., the management of the user area and the
system support area).
Assessing the Risk, Criticality, and Impact of Change
The persons responsible for managing the change process evaluate the feasibility and
impact of the proposed change in relation to their area of expertise and authority Areas
that should be included in the evaluation process are, as applicable: QA, regulatory
compliance, area management, the system owner/sponsor, the end-user, and technology
professionals (e.g., IT, engineering).
Their conclusions are then forwarded to the designated approver(s).
Specification of the Testing Requirements and Acceptance Criteria,
and Implementation of Change after Authorization
The approval or rejection of the proposed change by the identified responsible person(s)
must be documented. Areas to be included in the approval process are, as applicable: QA;
regulatory compliance; area management; the system owner/sponsor; the end-user; and
technology professionals (e.g., IT, engineering). The approver(s) are responsible for
evaluating the recommendations of the reviewers, making a decision on whether to
proceed with the proposed change, and initiating the implementation of the requested
Change management
99
change. The appropriate personnel must perform the change required in accordance with
the company policies, documented practices and/or system maintenance methodology.
Following the successful testing of the change, standard operating procedures are used
to:
1. control the transfer of the change into the operational environment
2. document a backout/rollback plan
3. verify functionality
4. document the final change
5. notify any impacted personnel
6. train any impacted personnel
During the move into the operational environment, any deviations and/or errors are
identified that cannot be corrected, the documented backout/rollback procedure must be
followed to restore the system to its premodified state.
Performing Applicable Regression Testing
All changes must be tested in the appropriate environment and documented according to
defined procedures that include:
a detailed plan for testing the change
the development of test script(s) for the new functionality, requirements, and/or
environment
the execution of the test(s)
the analysis and recording of the test results
Reviewing the Change(s) with an Independent Reviewer
Changes to computer systems must be reviewed in order to verify its compliance with the
change management process, to ensure that the appropriate procedures have been
followed, and that the validated status of the system has been maintained. The results of
the audit are recorded and saved as part of the change control file for the computer
system.
Updating the system and user documentation must always reflect the implemented
change(s).
Emergency change procedurals control must consider the specific requirements of
each computer system.
Provisions must be established for the management of an emergency change,
including expeditious documentation modifications.
Chapter 14
Training
Determination that persons who develop, maintain, or use electronic record/electronic
signature systems have the education, training, and experience to perform their assigned
tasks.
TRAINING IN THE REGULATED INDUSTRY

Personnel operating, maintaining, and programming computer systems should have
adequate training and experience to perform the assigned duties. One aspect to consider is
the extent of operation and system management. Training should include not only system
operation but also cover the significance of system faults (bugs), regulatory requirements,
system changes, security procedures, manual operation of the system, and documentation
of system errors. The organization must record the training of computer system related
personnel.
Training records are managed and retained by either the department providing the
training or by another designated department. The project leader for new systems, or the
system administrator for existing systems should coordinate the training. The system
1 Department of Health and Human Services, Food and Administration, 21 CFR Part 11,
Electronic Records; Electronic Signatures, Federal Register 62 (54), 1343013466, March 20,
1997.
21 CFR Part 11
102
administrator may maintain a copy of the training records throughout the life of the
computer system. Depending on the system, these records may include:
a description of the supplied training
a description of in-house training
copies of training manuals/materials
Any training conducted online must be performed in a controlled environment to ensure
that production systems and data are not adversely impacted. Appropriate security
measures must be taken to ensure that the training is isolated from the production
environment.
A linkage should be provided between training programs and the change management
process so that training requirements can be continually evaluated, adjusted, and
maintained in alignment with the current state of the operational computerized system.
Training manuals, handouts, personnel proficiency tests, and training procedures
supplied as part of the training provided, are an integral part of the systems
documentation. If these are available electronically, they can be referenced.
Training plans should be designed so that they are suitable for the intended audience.
As a result, there may be different courses for managers, supervisors, operations
personnel, system administrators, and other types of users.
Chapter 15
Security
Access to electronic records should be restricted and monitored by the systems software
through its log-on requirements, security procedures, and audit trail records. The
electronic records must not be altered, browsed, queried, or reported by external software
applications that do not gain entry through the protective system software. In addition to
the logical security built into the system, physical security must be provided to ensure
that access to computer systems and, consequently, to electronic records is prevented for
unauthorized personnel.
INTRODUCTION
To create a trusted digital environment within an enterprise, it must be ensured that both
the application and network components are truly secure. The FDA addresses the subject
of security for computer systems in the CGMP regulations and their associated policy
guidelines. Specifically, 21 CFR Part 211.68(b) and a recent guideline2 require that
appropriate controls are established over application and network components to ensure
that only authorized personnel can make changes to master production, control and/or
other required records. CPG 7132a.07, Inputs/Outputs Checking, contains specific
requirements to establish the necessary controls for records.
The first fundamental element of Part 11 is that it requires organizations to store

regulated electronic data in its electronic form, rather than keeping paper-based printouts
of the data on file, as had been the long-term practice in organizations performing
regulated operations. If information is not recorded on durable media, it cannot be
retrieved for future use. When retrievability is a requirement, then the procedural and
technological controls contained in Part 11 are essential to ensure data integrity. Only
1
2
FDA, Guidance for Industry: Computerized Systems Used In Clinical Trials, April 1999.
FDA, Guidance for Industry: Computerized Systems Used in Clinical Trials, April 1999.
21 CFR Part 11
104
regulated electronic records that meet Part 11 may be used to satisfy a predicate rule.
The second fundamental element in Part 11 is the high level of security required to
protect regulated electronic records. The requirements for application and network
component security can be found in Table 151. The controls that are implemented
resulting from the security related requirements are intended to build a trusted digital
environment. The attributes relevant to a trusted digital environment are:
private (secure information)
Table 151. Part 11 security related

requirements/controls.
Part 11 Description
PDA/ISPE Technological Controls3
11.10(c)
11.10(d)
Access controls
The system should be able to maintain electronic data over

periods of many years regardless of upgrades to the
software and operating system.
The system should restrict access in accordance with
preconfigured rules that can be maintained.
Any changes to the rules should be recorded.
The system should be capable of recording all electronic
record creation, update, and deletion perations. This
record should be secure from subsequent
unauthorized alteration,
The system should restrict the use of system functions
and features in accordance with preconfigured rules
that can be maintained. Any changes to the rules
should be recorded.
Where pharmaceutical organizations require that certain
devices act as sources of data or commands, the system
should enforce this requirement.
Not covered by the PDA/ISPE publication,
11.10(d) Authentication
11.10(e) Audit trail controls
11.10(e) Computer systems
time controls
11.10(g)
Authority checks
11.10(h)
Device checks
11.30
Technical controls
to open systems
Signature/record linking
11.70
11.100(a) Uniqueness of
electronic signatures
11.300
Electronic signatures
security
The system must provide a method for linking electronic

signatures, where used, to their respective electronic records,
in a way that prevents the signature from being removed,
copied, or changed in order to
falsify that or any other record.
The system should enforce uniqueness, prevent the
relocation of electronic signatures, and prevent the
deletion of information relating to the electronic
signature once it has been used.
The system should be able to identify changes to
electronic records in order to detect invalid or altered records,
ISPE and PDA, Good Practice and Compliance for Electronic Records and Electronic
Signatures, Part 2, Complying with 21 CFR Part 11, Electronic Records and Electronic Signatures,
September 1, 2000.
Security
105
authentic (proof of identity)

reliable (information integrity)
nonrepudiatable (undeniable proof of sender/receiver)
Another area of concern for the FDA relates to records sent over public networks, dial-up
connections, or public phone lines, accessed through external/internal web servers or by
database servers. A number of security issues need to be considered in order to keep these
records trustworthy. Intranets may face similar security issues if remote users connect to
the central network resources through a local link to an Internet, even when passwordprotected access for the users is provided to just a small portion of a private network.
These systems are considered by the agency as Open Systems (Part 11.30)4. For example,
the Internet provides a convenient medium to connect to other networks, but it does not
provide reliable security features, such as entity authentication, or protection from hostile
users or software.
In the USA, the Social Security Administrations policy5 is to integrate security into
the SLC for each application for the following reasons:
It is more effective. Meaningful security is easier to achieve when security issues are
considered as a part of the routine development process, and security safeguards are
integrated into the system during its design
It is less expensive. To retrofit security is generally more expensive than to integrate it
into an application during development
It is less obtrusive. When security safeguards are integral to a system, they are usually
easier to use and less visible to the user
Figure 151 depicts the types of security issues to consider in the internet/intranet world.
A vital factor necessary to achieve trustworthy records is to implement electronically
based solutions
4
5
FDA, Human Drug CGMP Notes, June 1999 (Vol. 7, No. 2).
US Social Security Administration, Systems Security Handbook, Release 5.0, Chapter 6.
21 CFR Part 11
106
Figure 151. Security issues to

consider.6
to strengthen the security. Trustworthy records must also be considered as part of the
requirements for record retention, archiving, and retrieval.
System security is also a concern for computer technology suppliers and contract
developers sites. The integrity of the computer environments that are used to maintain
the application software has to be verified when auditing suppliers and/or developers.
An assurance that the physical hardware, software, and the regulated electronic
records are maintained in a secure environment is critical to the validated status of a
computer system, particularly if it is an enterprise level system. Security must be
instituted at several levels. Procedural controls must govern the physical access to
computer systems (physical security). The access to individual computer system
platforms is controlled by network specific security procedures (network security).
Finally, application level security and associated authority checks control access to the
computer system applications (applications security).
PHYSICAL SECURITY
The following items are the key practices that are applicable for the security of the
physical environment of a computer system.
6
D.Coclin (Dean.Coclin@baltimore.com), Public Key Technology Overview, Courtesy of

Baltimore Technologies (www.baltimore.com).
Security
107
1. Limit physical computer access to authorized individuals.

2. The power sources for computer equipment must be controlled in order to prevent an
accidental uncontrolled loss of power, and fluctuations in voltage levels, which could
potentially damage equipment, software files and data files.
3. The power supplies for all computer equipment must meet both the manufacturers
requirements and safety national requirements for wiring, grounding, and radiation.
4. Critical environmental parameters that include, but are not limited to, relative humidity
and temperature, must be maintained within the computer equipment manufacturers
requirements. In cases where there are a number of requirements, the most stringent
should be utilized.
5. Critical environmental parameters must be monitored on a regular periodic schedule in
order to ensure that they continue to meet the manufacturers requirements.
6. The facility where the computer equipment resides must be equipped with the
appropriate fire-extinguishing systems or equipment.
7. The computer operations support, personnel facilities maintenance, and all other
personnel who need access to the physical environment must be responsible for
knowing where and how to obtain the applicable operational procedures,
understanding their contents, and for accurately performing their execution.
8. The use of authentication checks at physical access points in order to establish that the
machine-readable codes on tokens and PINs are assigned to the same individual.
These checks must be performed on the controlled access location(s) where the
computer hardware is resident.
9. The use of time-outs for under-utilized terminals in order to prevent unauthorized use
while unattended.
NETWORK SECURITY
The following items are the key practices that are applicable for the security of the
network environment.
21 CFR Part 11
108
1. Network resources have a qualified authentication mechanism for controlling access

(Part 11.10(d)). Table 152 shows the authentication requirements and the methods of
implementation. These authentication requirements and methods of implementation
are also applicable to applications security.
Table 152. Authentication.

Requirement
Implementation
The following features must be implemented:

automatic logoff
Automatic logoff,
biometrics
Unique user identification.
password
In addition, at least one of the other listed implementation PIN
features must be a procedure to corroborate that an
telephone callback
entity is who it claims to be.
token
unique user identification
Access control decision function are defined using access right lists,7 such as
Access Control Lists (ACLs), and these allow the allocation of use, read, write,
execute, delete, or create privileges. Access controls enable a system to be
designed in such way that a supervisor, for example, will be able to access
information on a group of employees, without everyone else on the network
having access to this information.
In the context of Part 11, access controls are an element of the authority checking
requirements.
2. The process for setting up user access to a network is the responsibility of the
appropriate network security administration personnel. The technical preparation,
education, and training for personnel performing administration tasks are fundamental.
The determination of what is required and the provision of documented evidence of
technical preparation, and the education/personnel training, are key regulatory
requirements (Part 11.10(i)).
7 In Windows 2000, an access control list (ACL) is a list of security protections that apply to an
entire object, a set of the objects properties, or an individual property of an object. Each Active
Directory object has two associated ACLs: The discretionary access control list (DACL) is a list of
user accounts, groups, and computers that are allowed (or denied) access to the object A DACL
consists of a list of access control entries (ACEs), where each ACE lists the permissions granted or
denied to the users, groups, or computers listed in the DACL. An ACE contains a security
identification with permission, such as Read access, Write access, or Full Control access. The
Security
109
system access control list (SACL) defines which events (such as file access) are audited for a user
or group.
3. Procedural controls that specify the manner in which network security is administered
must be established and documented. Network users must be trained in the policies,
practices, and procedures concerning network security.
4. The management of network user accounts is a key procedural control. This process
includes the request for the addition, modification, and removal of access privileges
(Part 11.10(g)). The request must be approved by the appropriate manager,
documented, and submitted to the network security administration for implementation.
5. There must be a procedure for granting controlled temporary network access for
personnel ((Part 11.10(d)).
6. In the event that a user leaves the company, there must be a process for notifying the
appropriate security administration as soon as the employee departs.
7. Provisions must be made for the regular monitoring of access to the network. There
must be an escalation procedure for defining the actions to be taken if unauthorized
access to the network is discovered.
8. A documented record of security administration activities must be retained.
9. Procedures must be established to control remote access to the network. Systems that
have connections to telephone communications through modems should have strict
access controls and restrictions. Access restrictions on these systems should be
designed to prevent unauthorized access or change. One possible method for
controlling remote access is telephone callback.
10. The use of time-stamped audit trails (Part 11.10(c) and (e)) to record changes, to
record all write-to-file operations, and to independently record the date and time of
any network system administrator actions or data entries/changes.
11. Unauthorized modification of the system clock must be prevented (Part 11.10(d)).
APPLICATIONS SECURITY
The following items are key practices that are applicable for the security of applications.
These practices are applicable to both networked and stand alone applications.
1. All applications must have a qualified authentication mechanism to control access

(Part 11.10(d)).
21 CFR Part 11
110
2. Software virus checking must take place periodically for the protection of the
application and data.
3. For each application, procedural controls must be established that specify the manner
in which application security is administered.
4. The process for setting up access to applications must be defined and executed by the
appropriate, application specific, security administration personnel. The technical
preparation, education and training for personnel performing administration tasks is
fundamental. The determination of what is required and the documented evidence of
the technical preparation, education and training of personnel is a key regulatory
requirement (Part 11.10(i)).
5. The management of the user application accounts is a key procedural control. This
process includes the request for the addition, modification, and removal of application
access privileges (Part 11.10(g)). The request must be approved by the appropriate
manager, documented, and submitted to the application security administration for
implementation.
6. There must be a procedure for granting controlled temporary application access for
personnel. As in the regular access, temporary access must follow security best
practices (Part 11.10(d)).
7. In the event that a user leaves the company, there must be a process for notifying the
appropriate security administration as soon as the employee departs.
8. There must be an escalation procedure for defining the actions to be taken if
unauthorized access to the application is discovered.
9. A documented record of security administration activities must be retained.
10. Procedures must be established to control remote modem access to applications.
11. If data input or instructions can only come from specific input devices (e.g.
instruments, terminals), the system must check the source of the input and the use of
the correct device should be verified by the operator (Part 11.10(h)).
12. The use of time-stamped audit trails (Part 11.10(c) and (e) to record changes, to
record all write-to-file operations, and to independently record the date and time of
any application specific operator actions or entries.
13. The use of time-stamped audit trails (Part 11.10(c) and (e) to keep track of any record
modifications carried out by the database administrator.
14. Use operational checks to enforce sequencing (Part 11.10(f)). Refer to Chapter 20 for
additional information.
15. The use of authority checks (Part 11.10(g)), when applicable, to determine if the
identified individual has been authorized to use the system, and has the necessary
access rights, operate a device, or to perform the operation required.
16. Unauthorized modification of the system clock must be prevented (Part 11.10(d)).
17. A shared computer system is a system used by various applications. Security
procedures should have been established regarding limiting access to confidential
information. In addition, access to data should be limited so that inadvertent or
unauthorized changes in data do not occur.
Security
111
OTHER KEY SECURITY ELEMENTS

Authentication
Three types of authentication are discussed in relation to Part 11: person authentication,
records/messages authentication, and entity authentication.
There are three person authentication methods: PIN and static passwords, PIN and
dynamic passwords, and biometric devices. Typically, the person authentication process
starts when the person enters a PIN into the system, and then his/her identity is
authenticated by providing a second piece of information that is known, or can be
produced only by the person (e.g., a password).
The most common methods for providing a strong authentication comprise of
automatic password generators (e.g., tokens) and smartcards. Tokens and smartcards
store information about the person and require the use of a reader device. To protect
against theft, the person must enter a password or PIN before the information in the token
or smartcard can be accessed.
Person Authentication
PINs and static passwords
Static passwords are still the simplest, most cost-effective and widely used
authentication mechanism. Although there are stronger mechanisms than
passwords (e.g. one-time passwords, challenge-response mechanisms,
cryptographic mechanisms, and biometric mechanisms) to achieve person
authentication, passwords are the most common mechanism, and often provide
adequate protection for computer resources.
Most systems today offer at least basic password management features. For the password
management features to be consistent with Part 11.300, they should include:
Password aging and expirationthis allows the password to have a lifetime, after
which it will expire and must be changed
Password historythis checks the new password to ensure that it has not been reused
during a specified amount of time or the specified number of password changes
Account lockingthis provides the automatic locking of an account if a person fails
to log in to the system after a specified number of attempts
Password complexity verificationthis performs a complexity check on the
password to ensure that it is complex enough to provide reasonable protection
against intruders who might want to break into the system by guessing the
password
Passwords file securityPasswords file shall be protected or otherwise secured so
that the password cannot be read by ordinary means. If the password file is not
protected by strong access controls, the file can be downloaded. Password files are
often protected with one-way encryption so that a plain-text password is not
21 CFR Part 11
112
available to system administrators8 or hackers (if they successfully bypass access

controls). Even if the file is encrypted, brute force can be used to learn passwords if
the file is downloaded (e.g., by encrypting English words and comparing them to
the file)
PIN and dynamic passwords
Passwords and PINs can provide a stronger means of authentication when
combined with other authentication techniques. Typically, the combination
involves something the person knows (e.g., passwords, PINs) and something that
the person possesses (e.g. tokens). Tokens are a software or hardware mechanism
that provides the person with a second piece of authenticating information.
Included within this group of mechanisms are:
One-time password generatorsAs the name implies, one-time passwords are
similar to traditional passwords since they are used in conjunction with a PIN, but
they are limited to one-time use. The advantage of this technique is that it prevents
the reuse of a compromised password. Commonly, a small hand-held device the
size of a credit card is synchronized with the target systems authentication scheme
and displays a one-time password that periodically changes. To access the target
system, the person enters an assigned PIN and password followed by the one-time
password currently displayed on the hand-held device. This method of
authentication provides additional security since the person must possess
knowledge of some personal information as well as the authentication token.
Challenge-response schemesThese look like one-time password generators and use
a similar synchronization mechanism; however, additional user actions are required
for authentication. It involves a challenge/response exchange with a new key being
used at each login.
8
One-way encryption algorithms only provide for the encryption of data. The resulting ciphertext
cannot be decrypted. When passwords are entered into the system, they are one-way encrypted, and
the result is compared with the stored ciphertext.
Security
113
Figure 15-2. SSL 3.0 Protocol.9

Certificate-based authenticationCertificate-based authentication is an increasingly
popular technique for strong authentication. An entity (see the definition later in this
chapter) wishing to be authenticated presents a digital certificate to the security
system. If the certified party is still trusted by the organization, then the server trusts
that the entity is who it claims to be. One benefit of certificate-based authentication is
that the entity does not have to have an established a relationship with, for example, a
server prior to being authenticated. This makes certificate-based authentication an
excellent solution for on-line commerce.
SSL 3.0 (Figure 152) protocol uses x.509v3 compliant digital certificates for
authentication. The most common usage of SSL is to authenticate a server to a
person, but it can also be used to authenticate a person to a server. By
authenticating a server, the person is assured of communicating with the correct
site. SSL also signs the information transmitted so that the person is assured that
the information has not been changed in transit.
A similar mechanism to certificate-based authentication is the authentication of
email transport. The authentication of email transport can be checked against the
identity of the recipient. The authentication of email users ensures that there is no
fraudulent user access to the system.
Biometric DevicesBiometric mechanisms tend to be best suited for the enforcement
of physical access, rather than for electronic access. That is, they work well where a
human guard (or proximity sensor, etc.) might be used, but they are not generally used
9

Baltimore Technologies (http://www.baltimore.com/).
for electronic access to a Web site. Consider, as an example, that the biometric
device in question (e.g., your fingerprint) is to be used for access to a Web site.
21 CFR Part 11
114
An attacker only has to capture your fingerprint data once, and they can then
replay it while masquerading as you. The following Internet address provides a
list of biometric devices:
http://www.icsa.net/services/consortia/cbdc/certified_products.shtml.
Records/Messages Authentication
Records must remain complete during transmission. An authentic record is one that is
proven to be what it purports to be and to have been created or sent by the person who
purports to have created and sent it. Records/messages authentication provides the means
to evaluate the integrity of the data/message. It involves technical and procedural controls
to detect unauthorized modifications to data/messages.
Unauthorized manipulation of records, audit trail records, and replay of transmissions
must be reliably identified as errors. The recipient can use three approaches to
authenticate data/messages.
encrypting records/messages using the recipients public key
signing the records/messages
employing a session key
The details of the above technologies are outside of the scope of this book. Additional
information on records/message authentication can be found in Chapter 22.
Entity Authentication
An entity can be a person, server, organization, or an account. One example of entity
authentication is Kerberos (see Internet address http://www.isi.edu/gost/info/Kerberos/).
Kerberos is an industry-standard authentication system suitable for distributed computing
over a public network. Implemented in Windows 2000, it uses public-key encryption
and hashing algorithms. Kerberos requires a Certificate Server that acts as a certification
authority and manages keys and digital certificates.10
This server maintains a database of secret keys for each entity, authenticates the
identity of an entity that wishes to access network resources, and generates session keys
when two entities wish to communicate securely.
The details of the above technologies are outside of the scope of this book.
Protection of Records and Audit Trails
All operations with records must be performed in a secure environment. This
requirement is applicable to:
10

Baltimore Technologies (http://www.baltimore.com/).
Security
115
records with/without signatures

records in transit
Organizations can use access controls, encryption (e.g., end-to-end data encryption)
and/or hashing to protect the integrity, authentication and privacy of records and/or to
prevent the loss of records. Records must be protected, for example, when confidential
patient data for specialist institutions is sent to distant places. Access controls were
discussed previously in this chapter.
Encryption can be used to protect records in transit between networks, including the
Internet. Tools such as network diagnostics or protocols analyzer can easily see
information as it is transmitted. This can be prevented by, for example, using SSLs, TLS,
or VPNs provide the necessary management of critical data across networks. These
solutions use encryption, digital signatures, and/or digital certificates to ensure data
privacy, identification of the originator of messages, and the verification of message
integrity.
End-to-end data encryption using PKI technology is implemented by encrypting
messages/records using the recipients public key, thus rendering messages/records
unreadable to anyone except the recipient. To encrypt messages/records the CA issues
distinctive digital signing certificates containing the public key assigned to the recipient.
Messages/records can only be decrypted using the recipients private key. The
combination of hashing and access controls provides another opportunity to protect
sensitive records.
An alternative method for sending records/messages via the Internet is using
Secure/Multipurpose Internet Mail Extensions (S/MIME). It is the standard for secure email within PKI. It enables secure communication between two or more parties by
ensuring the integrity and authenticity of the message given a digital signature, and the
confidentiality of the message if the message is encrypted.
Scheduled backups, established restoration procedures and off-site storage of media
are best practices that support recovery of records. The metadata associated with each
record must also be saved.
The protection of records by archiving must include taking into consideration the rate
of degradation of the electronic images, and the availability of the devices and software
needed to access the records.
Audit Trails
Audit trails refer to a journal that records modifications to electronic records. The persons
or automated processes operating on the users behalf may perform these modifications.
The audit-trail mechanism provides the capability to reconstruct the data prior to a
modification, and therefore must not obscure the previously recorded data. The tracking
mechanism must include a computer-generated time stamp that indicates the time of the
entry. Audit trails are computer generated and can be part of the record that has been
modified or a standalone record linked to the modified record.
21 CFR Part 11
116
The date and time of the audit trail must be synchronized to a trusted date and time
service. More information on this is provided in the next chapter.
One of the key controls for audit trails is the linking of the electronic record with the
audit trail. It must not be possible to modify audit trails; access rights to audit trial
information must be limited to print and/or read only. The combination of authentication,
digital certificates, encryption and ACLs, provides the technical mechanisms needed to
control the access to audit trail files.
Chapter 16
Source Code
REGULATORY GUIDANCE
We regard source code and its supporting documentation for application programs used
in drug process control to be part of master production and control records, within the
meaning of 21 CFR Parts 210 and 211. Accordingly, those sections of the current good
manufacturing practice regulations, which pertain to master production and control
records, will be applied to source code.1
When the end user cannot directly review the program source code or development
documentation (e.g., for most commercial off-the-shelf software, and for some contracted
software) more extensive functional testing might be warranted than when such
documentation is available to the user. More extensive functional testing might also be
warranted where general experience with a program is limited, or the software
performance is highly significant to data/record integrity and authenticity.2
INTRODUCTION
Source code is characterized as the human readable version of the list of instructions
(program) that cause a computer to perform a task. Source code in this context must be
viewed with a wide meaning since it includes, for example, the program listings for
custom-built applications; the configurable elements or scripts for configurable
applications; and the critical algorithms, parameters, and macro listings for off-the-shelf
applications. The source code typically includes, as part of the code, comments that
describe the subroutines or modules of the language used.
FDA, Compliance Policy Guide 7132a.15, Source Code for Process Control Application
Programs, April 1987.
2
21 CFR Part 11
118
The importance of the source code comes from the necessity to maintain the program and
to evaluate scripts, macros, or process specific parameters, and evaluating the impact of
any change on the rest of the program.
Up to the late 1980s, 80% of computer systems were specifically developed (i.e.,
custom/bespoke applications) as requested by the drug manufacturer and the source code
was a deliverable. By the early 1990s, the increased use of computers led to an increase
in the use of Standard Software Packages (see Chapter 10). Today, 80% of the software
utilized to supervise and control manufacturing processes is of the Standard Software
Packages type and only 20% is specifically developed (i.e., custom/bespoke
applications). Many of the specifically developed programs can be found, for example, in
programmable logic controllers. For canned configurable software applications, the
source code remains the proprietary information of the developer, so the likelihood of
acquiring it is very low.
Code inspections, reviews, and audits (see Chapter 9) consist of the smallest unit of
software designed to have a defined purpose during software development. These
debugging activities are necessary to verify that the code conforms to the software
requirements and the software development procedures. When testing the programs the
test cases are generated based on the internal code structure and include a 100% path
coverage. Code inspections and reviews are long and laborious processes performed by
the programmer and his/her peers; it is very difficult to perform code inspections and
reviews after the system has been installed.
Any serious software developer will follow a formal software development
methodology that includes the documentation necessary to support the software at each
development step. A quality audit of the software developer will determine the adequacy
of testing performed and whether the tests plan, procedures, and results are properly
documented. The quality audit must include an inspection of the unit testing process.
In 1996,3 the FDA provided an alternative to the source code review. When the source
code and the system specification cannot be obtained from the supplier, then black box
testing must be performed to confirm that the software meets the users needs and its
intended uses One way to make the source code available for the standard products of
configurable applications and for some contracted software is by establishing a thirdparty escrow account. This account enables organizations to ensure that application
software source code can be made available for inspection, if requested by a regulatory
agency. If an escrow account is implemented, organizations must adhere to the
confidentiality, copyright, and other legally binding agreements concerning the software
source code.
Before deciding on an escrow account, the legal department of the pharmaceutical
company must clearly communicate to the computer system supplier the source code
access requirements expected. Organizations should not require source code availability
for the following software: operating systems and system software, instruments and
controllers, and standard software packages.
3
Department of Health and Human Services, Food and Administration, 21 CFR Parts 808, 812,
and 820 Medical Devices; Current Good Manufacturing Practice (CGMP); Final Rule, Federal
Register, 61 (195), 5260252662, October 7, 1996, comment paragraph 136.
Source code
119
When the criticality and complexity (Refer to Appendix D) of the computer system
requires that standard product source code be made available, the legal department of the
pharmaceutical company should negotiate one of the following written agreements with
the software developer.
to deliver the source code to the organization
to place the source code into a third-party escrow account accessible to the organization
that the source code will be made available for inspection by a representative of a
regulatory agency
The agreement should also cover the updating of the source code held in escrow, and the
exact versions of which software applications must be held in escrow.
In any case, the software supplier or the developer must follow a documented change
management procedure, so that changes to source code are specified and documented.
Chapter 17
Hardware/Software Suppliers Qualification
It is the practice of organizations performing operations in a regulated environment to
work proactively with computer technology suppliers and contract developers, to ensure
that the system components conform to the quality requirements defined by each
organization.
Regulatory guidance
Systems should be in place to control changes and evaluate the extent of
revalidation that the changes would necessitate. The extent of revalidation will
depend upon the change's nature, scope, and potential impact on a validated
system and established operating conditions. Changes that cause the system to
operate outside of previously validated operating limits would be particularly
significant.
FDA, Guidance for the Industry: Computerized Systems Used In Clinical Trials,
April 1999.
When purchasing system software, standard instruments, microcontrollers, smart
instruments, and standard software packages, it is the practice of organizations to initially
rely on the history of the supplier and the industry operating experience as assurance of
the supplier capability of program performance suitability.
The history of the supplier, and the operating experience within the industry, can be
regarded as an assurance that the supplier is capable of providing a program that performs
in a suitable manner. However, the history of the supplier and the industry's operating
experience alone are not sufficient proof of a programs performance. The end-user is
responsible for ensuring that the program is suitable for its intended purpose when used
in a regulatory environment. Refer to Chapter 7 for additional validation activities.
When purchasing configurable software, it is often the practice to evaluate the
criticality of the application to the regulatory operation, in order to determine if a supplier
evaluation is warranted and feasible. If a supplier evaluation is not warranted or feasible,
the rationale for this must be documented and retained with the system validation
documentation. As in the previous case, the end user is responsible for ensuring that the
program is suitable for its intended purpose when used in a regulatory environment.
For custom-built application software, it is normal practice for organizations to
evaluate the developers quality system, including his procedural controls and experience,
in order to ensure that the system will conform to the quality requirements defined by the
organization.
The purpose of performing a hardware/software supplier qualification is to assess the
computing environments and technology products used to develop regulated applications.
Hardware/software suppliers qualification
121
These qualifications enable an objective evaluation of the supplier to be performed

and documented, including any findings, and to provide a baseline for computer system
validation.
The evaluation is accomplished through the inspection of:
practices used for the development, testing, maintenance, and use of the computing
technology
documented evidence that supports the integrity of the computing environment and the
technology products developed
historical data on the available operating experiences
During the qualification of computer technology suppliers and developers, a site visit and
inspection of the computing environment and related technology products is conducted.
Information is collected through formal inspection of the defined practices, verification of
supporting evidence, and interviews with key personnel. Information collected is
compared to the defined acceptance criteria and verified before completion of the on-site
visit.
Items to be considered are:
(a) Management
A defined organization is in place for the management of the project or contract
A quality assurance function is represented
The roles and responsibilities are clearly defined
(b) Documentation
Developers system specification exists and is linked with the organizations process/
operational requirements
Detailed hardware and/or software design specifications are documented
Hardware and/or software verification plans are documented
Test plans, procedures, scripts/specifications, input, and output data, and test reports are
available for inspection/audit
Verification of hardware and/or software functionality is documented
Verification of adherence to standards or development practices is documented
The source code is available, if applicable. Refer to Chapter 16
Site acceptance test plan and results are documented (if applicable)
User and technical support manuals and documentation are available
(c) Standards and procedures
Documented evidence is available that the hardware and/or software development
process is conducted under the auspices of documented standards, methodologies,
conventions, or procedures.
(d) Configuration management
A documented system is available and followed for configuration management.
Configuration management applies to: specifications, design and test documentation;
test plans and procedures; each item of hardware; development of software at each
21 CFR Part 11
122
incremental stage and for each constituent part/program and module; and related
documentation (e.g., manuals and reports).
Procedures for the security, archiving, and recovery of software products are clearly
defined.
(e) Change management
An existing and documented management of change process is followed during the
development and testing stages, and extending to the on-site installation and
commissioning activities, if required under the contract. This includes mechanisms for
requesting, submitting, evaluating, the impact, approving, implementing, testing,
releasing, and documenting the change. The traceability analysis can be used to
quickly identify areas impacted by the change.
The policies and procedures for the ongoing support and management of current and
obsolete products are clearly defined.
(f) Personnel qualifications
Documented evidence is available that establishes that the personnel designing,
programming, and maintaining the computer system have adequate training and
experience.
After visiting and inspecting the supplier, the information obtained is analyzed against
the acceptance criteria, current company requirements for the validation of computer
systems, and the regulatory expectations for computing environments. A report is
produced that documents the audit process, provides information on the current state of
validation of the supplier and what may be needed for compliance and, based on the audit
findings, provides a recommendation on whether the supplier can be used.
Any corrective actions that are required must be defined and implemented in order to
ensure that a validated state is achieved. The corrective actions must be scheduled and
follow-up evaluations planned for monitoring and tracking their progress.
Chapter 18
Maintaining the State of Validation
Once the system has been released for operation, the computer system maintenance
activities take over. The maintenance activities must be governed by the same procedures
followed during the development period.
The validated status of computer systems performing regulated operations is subject to

threat from changes in its operating environment, which may be either known or
unknown. Adherence to security, operational management, business continuity, change
management, periodic review, and decommissioning procedures and policies, provides a
high degree of assurance that the system is being maintained in a validated state. It is
essential that organizations have procedures in place to minimize the risk of computer
systems performing regulated operations in a nonvalidated state.
The maintenance of computer systems becomes an essential issue, particularly when a
new version of the supplier-supplied standard software is updated. A change control
procedure must be implemented whereby changes in the software and computer hardware
may be evaluated, approved, and installed.
When necessary, an additional analysis may be needed to evaluate the changes (e.g. an
impact analysis) on a computer system. The change control procedure should allow for
both planned and emergency changes to the system, and must include a provision for the
updating of the appropriate system documentation, including any procedures that may be
applicable. Records of changes to the system must be kept for the same period as any
other regulated production document.
Maintaining the state of validation
125
Table 181 1 summarizes the periods and events applicable during the operational life
of computer systems and the associated key practices.
SECURITY
Security is a key component for maintaining the trustworthiness of a computer system
and associated records. Security is an ongoing element to consider, subject to continuous
improvement. The security measurements that are implemented on a computer system
may be obsolete after several years.
Table 181. Period/Events Computer Systems

Operational Life.
Period/Event
Early Operational Life
Operational Life (Period)

Maturity
Aging
Representative characteristics
Phased rollouts
Corrective, adaptive, perfective,
and preventive maintenance
Maintainability issues of obsolete
technology (e.g., access to electronic
records)
Operational audit Periodic review
Performance
Re-engineering
evaluation
analysis
Problem reporting and

Key
Practices maintenance
In particular, after a system has been released for use, it should be constantly monitored
to uncover any security violations. One must follow up on any security violation, analyze
it and take proper action to avoid a reoccurrence.
Another important activity is the process to evaluate new security technologies and
their integration with computer systems. The procedural controls implemented as a result
of a risk analysis provide a starting point to look for technologies that can replace
procedural controls. These procedural controls are the result of security-related
implementation requirements identified during the risk analysis.
The security of computer systems is covered in Chapter 15.
Operational Management
Routine use of computer systems requires the following:
Training
All staff maintaining, operating, and using computer systems that perform
regulated operations must have documented evidence of training in their area of
1
G.J.Grigonis, E.J.Subak, M.L.Wyrick, Validation Key Practices for Computer Systems Used in
Regulated Operations, Pharmaceutical Technology, June 1997.
21 CFR Part 11
126
expertise. For users, the training will concentrate on the correct use of the
computer system, how to report any failure or deviation from normal operating
conditions and security. Refer to Chapter 14.
Hardware maintenance
The procedural controls applicable for the preventive maintenance and repair of
the hardware provide a mechanism for anticipating problems and, as a
consequence, possible loss of data. Modern hardware usually requires minimum
maintenance because electronic circuit boards, for example, are usually easily
replaced and cleaning may be limited to dust removal. Diagnostic software is
usually available from the supplier to check the performance of the computer
system and to isolate defective integrated circuits. Maintenance procedures should
be included in the organizations standard operating procedures. The availability
of spare parts and access to qualified service personnel are important for the
smooth operation of the maintenance program.
Organizations should use replacement parts that meet the design specifications of the
original computer system, or the system should be requalified to record that the
replacement part performs in accordance with the original specifications.
Computer systems used to control, monitor, or record functions that may be critical to
the safety of a product should be checked for accuracy at intervals of sufficient frequency
to provide assurance that the system is under control. If part of a computerized system
that controls a function critical to the safety of the product is found not to be accurate,
then the safety of the product back to the last known date that the equipment was accurate
must be determined.
Software maintenance
Maintenance of a software system includes
perfective maintenance or the correction of the system due to new requirements
and/or functions
adaptive maintenance, or the correction of the system due to a new environment that
could include new hardware, new sensors, or controlled devices, new operating
systems and/or new regulations
corrective maintenance, or the correction of the system due to the detection of errors
in the design, logic, or the programming of the system. It is important to recognize
that the longer a defect is present in the software before it is detected, the more
expensive it is to repair. Refer to Figure 22
preventive maintenance or the correction of the system to improve its future
maintainability or reliability, or to provide a better basis for future enhancements
As required by the regulations,2 all maintenance work must be performed after
the impact evaluation and approval of the work to be done. Refer to Chapter
13, Change Management.
For example, European Union (EU) GMP regulations, Annex 1111.
127
It is important to acknowledge that the maintenance of existing software can

account for over 70% of all effort expended by a software organization. Refer
to Figure 82. The computer hardware suppliers normally recommend
minimum maintenance schedules including accuracy checks for their
components.3
User support and problem management
These enable the reporting and registration of any problem encountered by the
users of the system. These can be filtered according to whether their cause lies
with the user or with the system itself, and fed back into the appropriate part of
the suppliers organization. Those problems that require a possible system change
are then managed through a change management procedure.
Archiving
All SLC documentation should be archived in an environmentally controlled
facility that is suitable for the material being archived, and that is both secure and,
where possible, protected from environmental hazards. A record of all archived
materials should be maintained.
Business Continuity
Business continuity procedures, including disaster recovery procedures, should ensure
minimal disruption in the case of loss of data or any part of the system. It is necessary to
ensure that the integrity of the data is not compromised during the return to normal
operation. At the lowest level, this may mean the accidental deletion of a single file, in
which case a procedure should be in place for restoring the most recently backed-up
copy. At the other extreme, a disaster such as a fire could result in loss of the entire
system. For this situation a procedure addressing the following should be in place:
specification of the minimum replacement hardware and software requirements to
obtain an operational system and their source
specification of the period within which the replacement system should be back in
production, based on business considerations
implementation of the replacement system
steps to revalidate the system to the required standard
steps to restore the archived data so that processing activities may be resumed as soon
as possible
The procedural control employed should be tested regularly and all relevant personnel
should be made aware of its existence and trained to use it. A copy of the procedure
should be maintained off-site.
3
FDA, Guide to Inspection of Computerized Systems in the Food Processing Industry, Office of
Regulatory Affairs (ORA), Division of Emergency and Investigational Operations (DEIO) and the
Center for Food Safety and Applied Nutrition (CFSAN),
http://www.fda.gov/ora/nspect_ref/igs/foodcomp.html.
21 CFR Part 11
128
Problem Reporting
The procedures to be followed if the system fails or breaks down must be defined and
validated. Any failures and remedial actions taken must be recorded. A procedure should
be established to record and analyze errors, and enable corrective action to be taken.
Refer to the section on computer system incidents located in Appendix E.
Control of Changes
Periodic Review
After being released for use, all computer systems must be periodically reviewed until the
system is no longer required for operation. The system documentation, system operation,
modifications, deviations, upgrades, and the electronic record management associated
with the software application must be reviewed to determine:
whether the corrective actions from the previous validation/audit were completed
the level of operational and performance conformance of the system
whether changes to the system have been properly tested and documented, and whether
the cumulative effects of small changes have caused a gradual degradation of system
performance and integrity
the impact of new/revised SOPs and regulations on the current CSV test methodology
the adequacy of calibration records and the continued suitability of the equipment to
which they apply
whether current operational SOPs are being followed
an audit of the current security measures that have been established for the system
the need for any refresher or update training for personnel
The system owner, system administrator, system user, and system support personnel must
participate in regular periodic reviews of the validated system. Upon completion of the
periodic review, the findings must be documented and an action plan developed to correct
any gaps found during the assessment.
A written certification, or other form of verification, must be provided to confirm that
the audit and inspection has been implemented, performed and documented, and that any
required corrective action has been taken. The certification should be filed in either the
system validation or project file.
Retirement
Refer to Appendix E.
129
Ongoing Verification Program

Computer system I/O errors can occur on validated systems. A computer component
(logic circuits, memory, microprocessor) or device (modems, displays), like mechanical
parts, can fail after they have been tested. Another source of computer systems I/O
malfunctions is electromagnetic interference (radio-frequency interference, electrostatic
discharge, and power disturbance). Software errors that were undetected during the
development process, may also be the source of I/O errors. In order to detect errors
during the operational life of a computer system before it can make decisions using
tainted data, an ongoing monitoring program must be established and followed to verify
the hardware and software I/Os. As part of the same ongoing monitoring program, stored
data should be checked for accessibility, durability, and accuracy.
Chapter 19
Part 11 Remediation Project
INTRODUCTION
As defined in CPG 7153.17, legacy systems are computer systems not in compliance with
Part 11. The first fundamental principle of Part 11 is that it requires organizations to store
regulated electronic data in its electronic form. The second fundamental principle is the
high level of security required to protect regulated electronic records.
Regulatory guidance
Firms should have a reasonable timetable for promptly modifying any systems not
in compliance (including legacy systems) to make them Part 11 compliant, and
should be able to demonstrate progress in implementing their timetable. FDA
expects that Part 11 requirements for procedural controls will already be in place.
FDA recognizes that technology based controls may take longer to install in older
systems.
FDA, CPG 7153.17, Enforcement Policy: 21 CFR Part 11; Electronic Records;
Electronic Signatures, May 1999.
As a result of the fundamental principles above, all computer systems performing
regulated operations must protect their electronic records by means of access controls,
audit trail controls, and Part 11-associated operational checks. For those computer
systems that implement electronic signatures, the security and control of electronic
signatures must also provided.
As part of the remediation process for legacy systems, a retrospective evaluation
should be performed on the electronic records recorded after August 1977 using the
legacy systems. The objective is to assess the complete system operation, including:
procedural controls, change control, data and system security, any additional
development or modification of the software using an SLC approach, the maintenance of
data integrity, system backups and operator (user) training.
If evaluation results do not meet the current regulatory requirements, then the
retrospective evaluation would not in itself support the trustworthiness of the electronic
records. This is an area of concern for electronic records recorded using legacy systems
after August 1997.
Figure 191 illustrates a complete Part 11 Remediation Project. Like any project, the
schedule is based on priorities, time, and the availability of resources.
21 CFR Part 11
132
Figure 191. Complete Part 11

Remediation Project.
This chapter identifies the key principles for remediating a noncompliant Part 11
system, but it is not intended to cover everything that an organizations management
needs to do in order to achieve and to maintain compliance with Part 11. The
PDA/GAMP Forum document1 is an excellent resource for planning a Part 11
Remediation Project.
EVALUATION OF SYSTEMS
The evaluation of computer systems performing regulated operations is the first phase to
achieving an organized, prioritized, and balanced Part 11 Remediation Project approach.
The objective of the evaluation is to identify the systems functional and/or procedural
gaps; results of the evaluation will determine whether the operational, maintenance, or
security procedures specific to the system will provide a controlled environment, which
ensures the integrity of the electronic records and/or signatures as stated in the Part 11
requirements.
An evaluation plan is needed in order to define the nature, extent, schedule, and
responsibilities of the evaluation and process.
Each system performing a regulated operation must be identified and the operation it
performs must be well understood in order to prioritize the work. Data and process flow
diagrams may be used as tools for reviewing the operation. Performing Step 2 can obtain
the priority rating that is applicable for each system in the Criticality and Complexity
Assessment located in Appendix D. Other factors to take into account during the
prioritization process are the components and functions that have regulatory implications.
After the computer systems have been prioritized, they are evaluated by following a
worksheet similar to that provided in Appendix G. The Part 11 issues to be evaluated are:
2. Security
a. System security
c. Access codes and password maintenance
Access codes and password security
Password assignment
1
PDA/GAMP Good Practice and Compliance for Electronic Records and Signature, Part 2,
Complying with 21 CFR Part 11, Electronic Records and Electronic Signatures, Version 1,
September 2001. Published jointly by PDA and ISPE.
Part 11 remediation project
133
4. Audit Trails
a. Audit mechanism
b. Metadata
a. Electronic-signature without biometric/behavioral controls
b. Electronic-signature with biometric/behavioral controls
An evaluation report must be generated for each computer system, which summarize the
current operation of the computer system, allocates its priority, provides a reference to
any supporting documentation, and identifies the compliance gaps in the system. Based
on the information in the evaluation reports, a Corrective Action Plan can be generated.
CORRECTIVE ACTION PLANNING

The purpose of the Corrective Action Plan is to define the overall activities, schedule,
costs, and responsibilities necessary to guide the development and implementation of
technological and procedural controls to bring the systems into compliance with Part 11.
The plan should identify any existing technological/procedural controls that may be
modified or new technological/procedural controls that need to be implemented in order
to ensure that the regulatory requirements are completed in a consistent and uniform
manner. The remediation action items identified in the Part 11 assessment should be
documented in a detailed implementation plan.
Using the evaluation reports, the remediation activities, available resources, and
project schedule, the business cost of the remediation approach can be estimated. This
will enable a business decision to be made regarding the remediation or replacement of
the current system based on the cost-effectiveness of the system and its operational
feasibility.
The plan must contain the activities that place emphasis on achieving consistent, high
quality, sustainable compliance solutions. The remediation process consists of six major
activities. These activities are:
interpretation
training
remediation execution
new applications assessments
21 CFR Part 11
134
application upgrade assessments

supplier qualification program
Once the corrective action plan has been approved, it can then be executed.
REMEDIATION
During the remediation phase, the computer systems are brought into compliance by
implementing the procedural and technological controls determined by the Corrective
Action Plan. In addition, the processes needed to sustain the compliance solutions are
implemented.
Interpretation
The ability to evaluate the risks of a particular system in relation to 21 CFR Part 11
requires athorough understanding of the regulation and a consistent interpretation. The
objective of theinterpretation phase of this plan is to provide current, consistent,
regulatory interpretation of 21CFR Part 11 to IT and interested outside parties.
Training
In the context of Part 11, awareness and understanding of the regulation is fundamental to
the success of the Part 11 Remediation Plan. The objective of the Part 11 training is to
ensure that all system and technical owners have an appropriate level of knowledge of
Part 11.
Other forms of training need to be developed in order to support the maintenance of
regulatory compliance of computer systems. Refer to Chapter 14 for more information.
Remediation Execution
Once the Corrective Action Plan is approved, the computer technology suppliers and
developers are requested to identify how the deficiencies can be overcome. When
appropriate, procedural controls need to be developed in order to address the deficiencies
that cannot be solved by technological controls.
It is probable that the implementation of technological controls will require a
comprehensive SLC including: the recommendation, conceptualization, and
implementation of the new technology; the release and early operation of the new
technology; and the decommissioning and disposal of the old technology. If the
technological implementation fails, the failure should be documented along with details
of the corrective action taken. Once this action has been taken, the system must be
reevaluated in the same way as any other system that had been subject to an upgrade or
correction.
When all of the action items applicable to a computer system have been implemented,
it can be formally released for operation, and for support under a maintenance agreement.
Part 11 remediation project
135
The Corrective Action Plan should be periodically reviewed because the evolving
technology requirements will need to be considered and the plan revised accordingly.
New Applications and Application Upgrade Assessments
The objective of Part 11 assessments for new applications, and for application upgrades,
is to identify the Part 11 gaps before releasing the system into production. All gaps must
be managed using procedural and/ or technological controls.
New systems and upgrades to systems that are released into production must have a
high level of compliance. However, due to limitations on the technology used by
suppliers, it is not always possible to implement systems that are fully compliant.
These systems must be assessed, the Part 11 gaps recorded, and a formal plan
established to remediate the gaps.
The project team is responsible for completing this assessment.
Suppliers Qualification Program
A key business strategy has been the outsourcing of work to computer technology
suppliers and developers. The objective of qualifying computer technology suppliers and
developers is to evaluate and monitor these strategic partners for Part 11 compliance
and to provide an input to the partner selection and partner relationship management
processes.
For each qualification performed, a report must be prepared that describes the results
of the qualification.
REMEDIATION PROJECT REPORT

The Remediation Project Report provides evidence of successful project completion. It
must describe the technological and procedural controls and associated activities
necessary to make the computer technologies that were assessed compliant with Part 11.
This report and all supporting documentation should be archived.
Once a legacy system has achieved a satisfactory, documented Part 11 compliant state,
any subsequent changes can be prospectively validated.
Chapter 20
Operational Checks
The objective of operational checks is to enforce the sequencing of steps and events as
appropriate.Regulatory The application-dependent algorithms consisting of sequencing of
operations, instructions to the operator, critical embedded requirements, and safety
related precautions, are encompassed in the computer program(s) driving the computer
systems. These application-dependent algorithms and Part 11 requirements become part
of the computer systems validation model (refer to Chapter 4 for more information)
through the operation process input refer to Figure 41.
Regulatory guidance
Use of operational system checks to enforce permitted sequencing of steps and events,
as appropriate.
Department of Health and Human Services, Food and Administration, 21 CFR Part
11, Electronic Records; Electronic Signatures, Federal Register 62 (54), 13430
13466, March 20, 1997.
The lack of operational checks to enforce event sequencing is significant if an operator
deviates from the prescribed order of computer system operation steps, resulting in an
adulterated or misbranded product, and/or data integrity.
There are three types of operational checks: instructions to operators, system
sequencing, and Part 11 operational checks.
INSTRUCTIONS TO OPERATORS
At the operator level, the purpose of operational checks is to guide the operator during
specific actions. The intention is to prevent operators from operating the system and/or
signing records outside of the preestablished order. The preestablished order is specified
in the system requirements specification. Two examples of typical operator actions that
require the enforcement of a sequenced approach include:
Signing a batch record to show that the addition of an ingredient occurred before the
ingredient was actually added
Verification by a second person that an activity was performed as specified by the batch
record.
Operational checks
137
OPERATION SEQUENCING
Operational checks are normally presenting process control computer systems. These
systems may contain code that is part of the master production record. At the system
level, the purpose of operational checks is to execute algorithms, sequencing of
operations, and safety-related functions as required in the applicable customer
specification. Inspections and testing are fundamental processes to be performed during
the validation of critical system sequences. In addition, an ongoing program must be
established to frequently verify that critical operations occur in the proper sequence.
Section 211.188(b)(11) of the Current Good Manufacturing Practice Regulations is an
example of operation sequencing. This section requires that batch production and control
records include documentation that each significant step in the manufacture, processing,
packing, or holding of a batch was accomplished, including identification of the persons
performing, directly supervising or checking each significant step in the operation.
When the significant steps in the manufacturing, processing, packing, or holding of a
batch are performed, supervised, or checked by a computerized system, an acceptable
means of complying with the identification requirements of 21 CFR 211.188(b)(11)
would consist of conformance to all of the following1:
documentation that the computer program controlling step execution contains adequate
checks, and documentation of the performance of the program itself
validation of the performance of the computer program controlling the execution of the
steps
recording specific checks in batch production and control records of the initial step, any
branching steps and the final step
In general, assessment to the compliance to 211.188(b)(11) 1) includes documented
demonstration that the computer system examines the same conditions that a human
being would look for, and that the degree of accuracy in the examination is at least
equivalent.
Other examples of typical system sequencing include the correct automatic operation
of equipment controlled by a computer system such as fluid bed system, automated
packing lines, and product release is another function that involves sequencing.
PART 11-RELATED OPERATIONAL CHECKS

Part 11 contains specific sequencing requirements that must be implemented using
computer technologies. These operational checks usually fall in the category of operation
sequencing and are built into the software/hardware. To highlight the importance of these
operational checks, the author suggests that specific categories are allocated for them as
follows:
1
CPG 7132a.08, Identification of Persons on Batch Production and Controls Records, November
1982.
21 CFR Part 11
138
Electronic signature manifestation, 11.50(b)

The electronic signature must be displayed in any human readable form, including
printouts and video displays:
immediately after the signature is executed
after displaying a signed record
when printing signed electronic record(s)
The printed name of the signer, the date and time of signing, and the meaning
associated with the signing must be displayed.
Multisigning, 11.200(a)(1)(i)
When an individual executes one or more signings that are not performed during
a single, continuous period of controlled system access, each signing must be
executed according to the following:
1st signing: using both the userID and the password components
2nd + and subsequent signings during a period of continuous, controlled access:
either re-entry of the password alone, or use of both the userID and password
components
One signature can apply to multiple data entries on a screen as long the items that
the signature applies to be clearly indicated
Flagging unauthorized use of passwords and userID codes, 11.300
Systems that use electronic signatures must be designed so that unauthorized
attempts to use the signature are detected and reported to security management.
The security system should be capable of identifying situations where misuse
occurs and should notify security management appropriately. Repeated attempts
at unauthorized access must result in access to the software application being
automatically disabled. A notification option may be used to log the message to a
historical file or to send a message to a system administrator workstation. A
process for investigating attempted security violations must be defined so that
security violations are handled promptly.
Authority checks
The system must implement authority checks to ensure that only authorized
individuals can use the system to sign records, access the operation or device,
alter records, or perform the operation at hand, 11.10(g). These authority checks
are based upon the various roles and responsibilities assigned to individuals who
are known to the system. The computer system must be designed to make
distinctions between the control access to the system, to functions in the system,
and to input and output devices used by the system.
The system shall be designed to implement device checks (11.10(h)) including
recording the location (node) of the workstation where each entry was made. Device
checks enable the software application to determine whether the input being generated
by a particular device is appropriate. The use of the term as appropriate does not
require that device checks are performed in all cases. These checks may be used when
certain devices have been selected to be legitimate sources of data input or commands.
Operational checks
139
For example, in a network environment it may be necessary for security reasons to

limit the issuance of critical commands to a particular authorized workstation.
The application must detect when a workstation has experiences a long idle period and
must be able to automatically log out the user.
Audit trails refer to a journal or record of modifications to electronic records by the
users or by processes operating on the users behalf for tracking purpose. Data must be
protected from unauthorized modification and destruction to enable the detection, and
after-the-fact investigations, of security violations.
This mechanism provides the capability for modified data to be reconstructed in
its previous form and therefore the modification must not obscure the previously
entered data. The tracking mechanism must include a computer-generated time
stamp that indicates the time of the entry that modified the record and the types of
modifications performed. The requirements for audit trails are as follows:
they must be computer generated
they can be either part of the electronic record itself or a separate record
they must not be modified by the person who create them
they must indicate, 11.10(e):
when the data was first entered and by whom
when who made to the data and any changes
Signature/record linkage must be computer generated, 11.70
The linkage may be achieved by linking the userID, which could be obtained
from a password/security file. The signatures must not be capable of being
removed, copied, changed, or transferred in order to falsify an electronic record
by ordinary means. The signer's full name doesn't have to be embedded in the
record itself. The field containing the name can be a pointer that points to a file
containing the full names of the signer.
The link must be retained for as long as the record is kept, just as a handwritten
signature stays with the paper, long after an employee has departed a company.
Although a userid/password can be removed from a current user database, it must
still be retained in an archive in order to maintain the signature and record
linkage.
VALIDATION OF OPERATIONAL CHECKS2

An acceptable method for the validation operational checks would be as follows:
1. The documentation of the program, including a requirements specification that
describes what the software is intended to do
2. The performance of inspections and testing so that no step or specification can be
missed or poorly executed/assigned
3. Documentation of the initial and final steps
2
FDA, Guide to Inspection of Computerized Systems in Drug Processing, February 1983.
Chapter 21
Compliance Policy Guide (CPG) 7153.17
INTRODUCTION
The main objective of Compliance Policy Guide (CPG) 7153.17 is to provide specific
guidance to the FDA field investigators on whether or not to pursue regulatory action
against an organization under Part 11. It establishes the direction to be followed by these
investigators when they find systems that are not properly compliant with Part 11. CPG
7153.17 makes the following points:
Regulatory guidance
FDA will consider regulatory action with respect to Part 11 when the electronic
records or electronic signatures are unacceptable substitutes for paper records or
handwritten signatures, and that therefore, requirements of the applicable
regulations (e.g. CGMP and GLP regulations) are not met. Regulatory citations
should reference such predicate regulations in addition to Part 11.
FDA, CPG 7153.17, Enforcement Policy: 21 CFR Part 11; Electronic Records;
Electronic Signatures, 5/13/99.
The FDA defines legacy systems as those systems in place before August 20, 1997.1
The main area of enforcement of this CPG comes from the no grandfathering clause.
This clause applies to legacy systems and the electronic records stored by these
systems. The no grandfathering clause indicates that Part 11 provisions relating to
record creation do not cover electronic records that are created before the effective
date of the rule. These records would not need to be retrospectively altered.
The four elements to be considered by the field inspectors when pursuing regulatory
actions are
nature/extent of the deviation
impact on product quality/data integrity
adequacy/timeliness of the corrective action plan
compliance history
FDA, Human Drug CGMP Notes, Vol. 7, No. 3, September 1999.
Compliance policy guide (CPG) 7153.17
141
Nature/Extent of the Deviation

FDA inspections are conducted for many reasons and the extent of compliance of
computer systems with Part 11 may be evaluated during any of these inspections. If
applicable, a field investigator will review electronic records in order to evaluate their
level of compliance with the predicate rule. At the same time, the investigator may check
adequacy of record keeping arrangements for compliance with Part 11.
Some of the elements to be evaluated are:
Does the system meet the appropriate Part 11 requirements?
Has the system been validated?
Have Part 11 procedural controls been met?
Has the company submitted certification for electronic signatures?
Audit trails are the most important technical consideration and the most important
element of Part 11. These are generated when data, which is stored on durable media, is
modified. Audit trails ensure that changes to the electronic records do not obscure the
source data. Electronic data as well as data collected manually can be amended and the
amendment must be recorded and justified2. Changes to electronic records will always
require an audit trial3; these changes must not obscure the original record. Audit trails
must include the original entry, the new entry, the reason for the change, the date and the
time of the change, and the electronic signature.
Audit trails may be a special consideration during a recorkeeping inspection. The
requirements for audit trails are as follows:
They must contain who did what, wrote what, and when (date and time stamp)
They can be part of the electronic record or a record by itself
They must be computer generated and saved electronically in chronological order
They must be created when the operator performs entries and actions that create,
modify, or delete electronic records, e.g., command to acknowledge an alarm
They must not be modified by the person who creates them
Impact on Product Quality/Data Integrity
This element of guidance to be used by the FDA inspectors is based on considerations
such as the GxP criticality of the computer system.
The criticality, or product risk, must be assessed for both the direct and indirect
impact of the computer system on the product. The extent of the validation/qualification
activities to be carried out will depend on the results of this assessment. The criticality is
based on the potential impact on the product, and existing data does not justify a lower
rating.
2
Final Rule for 21 CFR Part 211.100 (1996), Current Good Manufacturing Practice For Finished
Pharmaceuticals.
3
21 CFR 11.10(e).
21 CFR Part 11
142
Other factors that should be considered to assess product impact are:

manufacturing
labeling
packaging
master batch records/device history records
quality records (the data collected and used to make product release decisions)
product performance (product and manufacturing)
delivery
inventory
The electronic data obtained from the areas mentioned above are highly critical. The field
inspector will look for evidence of any lack of control associated with the electronic
records.
Adequacy/Timeliness of the Corrective Action Plan

The corrective action plan may be contained within a validation plan. The validation plan
is a document that describes the companys overall philosophy, intentions, timetable, and
approach to be followed for the corrective action compliance program. The approach to
be taken may consist of writing procedures, performing verifications, and/or performing
qualification activities. Refer to Chapter 19 for information on remediation project.
The FDA expects procedural controls relating to Part 11 to be in place. These include
Change Control, Security, System Start-Up/Shutdown, Data Collection and Management,
Backups, System Maintenance, System Recovery, Training, Periodic Reviews,
Revalidation, Contingency Planning, Performance Monitoring, and Decommissioning.
These procedural controls must be verified during the qualification of the computer
system and periodically evaluated.
Compliance History
The compliance history of the company being inspected, especially with regard to data
integrity, will be taken into account in order to intensify their scrutiny of electronic
records. If a company has a history of Part 11 violations or inadequate or unreliable
record keeping, this would make any additional Part 11 deviations more significant. Until
firms reach full compliance, FDA investigators will be more vigilant when attempting to
detect inconsistencies, and unauthorized changes.
Chapter 22
Electronic Records
Electronic record means any combination of text, graphics, data, audio, pictorial, or
other information representation in digital form that is created, modified, maintained,
archived, retrieved, or distributed by a computer system. 21 CFR Part 11.3(6).
WHAT CONSTITUTES AN ELECTRONIC RECORD?

An electronic record is any information that is recorded in a form that:
requires a computer or other machine to process the electronic record2 and that satisfies
the legal definition of a record in 44 U.S.C. 33013
only a computer can process4 and that satisfies the definition of a record in 44 U.S.C.
3301
FDA, Electronic Records; Electronic Signatures Final Rule, 62 Federal Register 13430 (March
20, 1997).
2
US Department of Defense, DOD 5015.2-STDDesign Criteria for Electronic Records
Management Software Applications, November 1997.
3
3 Records includes all books, papers, maps, photographs, machine readable materials, or other
documentary materials, regardless of physical form or characteristics, made or received by an
agency of the United States Government under Federal law or in connection with the transaction of
public business and preserved or appropriate for preservation by that agency or its legitimate
successor as evidence of the organization, functions, policies, decisions, procedures, operations, or
other activities of the Government or because of the informational value of data in them. Library
and museum material made or acquired and preserved solely for reference or exhibition purposes,
extra copies of documents preserved only for convenience of reference, and stocks of publications
and of processed documents are not included.
4
36 CFR 1234.2.
Electronic records
145
The definitions above are consistent with Part 11. Per Part 11, paragraphs 22, 45, and
72 in the preamble refer to when data on transient memory become electronic records. In
summary, the primary attribute to take into consideration for the above definitions is
record retrievability. It is essential to take this into account when deciding the
applicability of Part 11.
The current federal legislation, Electronic Signatures in Global and National
Commerce Act, defines electronic records in more general terms than Part 11.
For this definition, the primary attribute to take into consideration is whether the data
is accessible once it is put into storage, not the retrievability of the record. This
definition would consider data in transient memory to be an electronic record.
WHAT CONSTITUTES A PART 11 REQUIRED RECORD?

Part 11 required records are those:
that are defined in the applicable predicate rule, especially those used to determine
product quality
that may not be explicitly identified in the applicable predicate rule, but is a requirement
that the organization imposes upon itself
One example of this self-imposition by an organization might be the records required
by procedural controls such as those for master batch records.
HOW SHOULD PART 11 RECORDS BE MANAGED?

Required electronic records that meet Part 11 are used to satisfy the applicable predicate
rule. The following are some basic principles that must be adhered to:
Procedural controls must be established and maintained providing instructions for
identification, collection, indexing, filing, storage, retention requirements,
maintenance, and disposition of quality system records. (Section 5.1, FDA, draft of
Guidance for Industry: 21 CFR Part 11; Electronic Records; Electronic Signatures:
Maintenance of Electronic Records, July 2002)
Computer environments that contain and manage such records must be validated. (Part
11 and EU GMPs Annex 112)
Detailed procedures relating to the system in use should be available and the accuracy
of the records should be checked. (EU PIC/S and Section 5.3, FDA, draft of Guidance
for Industry: 21 CFR Part 11; Electronic Records; Electronic Signatures:
Maintenance of Electronic Records, July 2002)
After the computer environment has been validated, there should be an ongoing
monitoring program to verify the data/control/ monitoring interface(s) between the
system and external equipment in order to ensure the correct input and output
transmission of data (CPG 7132a.07)
As part of an ongoing monitoring program, stored data should be checked for
accessibility, durability, and accuracy (EU GMPs Annex 1113)
21 CFR Part 11
146
Electronic records should be secured by physical or electronic means against wilful or

accidental damage. Electronic records should only be entered or amended by persons
authorized to do so (EU GMPs Annex 1113, 8)
Data should be protected by backing-up at regular intervals (EU GMPs Annex 1114)
There should be available adequate alternative arrangements for systems that need to be
operated in case of a breakdown. The procedures to be followed if the system fails or
breaks down should be defined and validated (EU GMPs Annex 115, 16)
Any alteration to a required record should be recorded with the reason for the change
(EU GMPs Annex 1110)
There should be a defined procedure for the issue, cancellation, and alteration of
authorization to enter and amend data, including the changing of personal passwords
(EU GMPs Annex 118)
If changes are proposed to the computer equipment or its programs, the checks
mentioned above should be performed at a frequency appropriate to the storage
medium being used
It is particularly important that the electronic records are readily available throughout
the period of retention (EU PIC/S and Section 5.2, FDA, draft of Guidance for
Industry: 21 CFR Part 11; Electronic Records; Electronic Signatures: Maintenance of
Electronic Records, July 2002)
MINIMUM RECORD RETENTION REQUIREMENTS

Part 11 defines how electronic records must be managed. The retention requirements of
such records are contained in the applicable predicate regulation. Table 221 refers to
some of the sections in the predicate regulations or guidelines that specify record
retention requirements.
Table 221. Records Retention Requirements.

Area Section
GLP
GLP
GLP
GLP
GLP
GLP
GLP
GLP
GLP
GLP
GCP
GCP
GMP
GMP
GMP
Regulation
Organization and personnel records

FDA 58.3
Specimen and data storage facilities
FDA 58.51
Storage and retrieval of records and data
FDA 58.190 & 58.195
Organization and personnel
OECD 1.1 & 1.2
Sampling and storage
OECD 6.1
Archive facilities
OECD 3.4
Characterization
OECD 6.2
SOPs
OECD 7.2
Storage and retrieval
OECD 10.1 & 10.2
Storage and retrieval with archive facilities
OECD 10.2
Supply and handling of investigational product(s) ICH 5.14
Trial management, data handling, record keeping ICH 5.5
Records and documentation
FDA 211.180
Personnel records
EU PIC 2
Documentation
EU PIC 4
Electronic records
147
The organization owning the electronic records may impose more stringent retention
requirements that may be based on legal requirements.
WHEN ARE AUDIT TRAILS APPLICABLE FOR ELECTRONIC

RECORDS?
The point at which data becomes an electronic record should reflect its origin and
intended purpose. For manufacturing and laboratory records this may be the initial point
of data acquisition but for a clinical trail report it may be from the reports first official
use. Electronic records can be divided into three categories based on their intended
purposes: instructions, events, and reviews. Using these categories, it is possible to
determine when audit trails are applicable. Table 222 summarizes these concerns and
needs.
INSTRUCTIONS
The instructions embedded in computer program(s) either provide information on what
the operators are supposed to do, or information on how the equipment or processes are
intended to function. For example, for manufacturing operations, the batch production
and control records may provide procedures, controls, instructions, specifications, and
precautions to be followed when using computer systems. These programs may also
contain control data for product formulation, batch size, yields, and automated in-process
sampling/testing procedures.
Instructions may be modified before being loaded into the associated workstation. The
modification may be related to the adaptation of a process or the introduction of a new set
of instruction to the operator. In either case, the modification must be qualified and the
operators trained to use it.
Instructions are version controlled and need to go through a formal approval process
before they can be used. When working with instructions in electronic format, the audit
trails should start when the instruction is official approved.
In a paper-based environment, if a procedural control is, in fact, a draft, (e.g., an
early version of a procedural control that is ultimately not adopted) the applicable
predicate rule would probably not require that it be retained. The same approach is also
applicable to instructions in electronic format.
When there is a deviation from an instruction, the system must have the capability to
record the deviation and the associated audit trail must contain the action(s) carried out
by the operator. In addition, an investigation should be carried out and the cause of the
deviation justified.
21 CFR Part 11
148
EVENTS
Events are the various defined actions or activities performed by personnel, equipment, or
instruments. The records of events provide evidence that make it possible to establish
what
Table 22-2. E-records Concerns and Needs.

Record Type
Key Concern
Instruction records (records that

Is this the latest
establish or define expectations,
approved version?
e.g., Validation Plans, Protocols,
What version was
SOPs, specifications, recipe, method) used for that specific
batch or experiment?
Key Control
version and
distribution
management
control of change
access security
audit trail
Event records (the information is

Who? When? What?
recorded contemporaneously with
Reason of changes
the actions, e.g., electronic data
or corrections?
from laboratory instruments or
manufacturing control operations)
Review records (records that provide Who takes responsibility? signature requirements
evidence that events, information,
Responsibility for what? record/signature
or data have been evaluated for
link
Review, approval,?
content, completeness, or accuracy, Based on what data?
e.g., reviews required in 21 1.192.
Usually these records are approved.)
actually happened during a process. These types of documentation cover important

aspects that need to be recorded in order to demonstrate that systems performing
regulated operations are adequately controlled. For example, for a computerized drug
process, certain information required by the CGMP regulations to be included in a master
production record is contained within the source code for the application program.
For process equipment controlled by computers, the events are recorded at the time of
occurrence by the computer systems. Instrumentation connected to computer systems
may measure process parameters or document actions such as addition of ingredients as a
specific process event. The recording of processing events and operator actions must be
audit trailed immediately, and should not wait for approval.
As a consequence of the distinctiveness of electronic records resulting from events,
one area of awareness is amendment and (logical) deletion of these electronic records.
Procedural controls must define the process of correction. The person supervising or
checking the operation should perform amendments to event files.
Electronic records
149
REVIEWS
Reviews are written evidence that events, instructions, information, or data have been
evaluated for content, completeness, or accuracy. Critical reviews are those where
acceptance criteria are compared against the results of a process to ensure that the data
conforms to requirements.
An example of a required review is the one mandated by Section 211.188(b)(11) of the
CGMP regulations. It is a requirement that batch production and control records include
documentation that verified that each significant step in the manufacture, processing,
packing, or holding of a batch was accomplished, including the identification of the
persons performing, directly supervising, or reviewing each significant step in the
operation. The intention of the regulation is to ensure that each significant event in a
process is reviewed and that there is record evidence to show this. It is quite possible that
a computer system could achieve either the same or higher degree of review assurance. In
this case, it may not be necessary to specifically record the checks made on each series of
steps in the operation.
Reports may be considered to be reviews. For electronic documents that provide an
interpretation of data drawn from electronic records that are already subject to control,
there is no clear need to reimpose these controls until the document is used for some
official purpose. In a case of submissions, once a submission has been assembled and
approved ready for submission to a regulatory agency it would then become an electronic
record subject to audit trails.
If any electronic record resulting from an event, instructions, information, or data that
was part of a review is modified, the approver and reviewer must be prompted to repeat
the review of the data. This repeat review must be prompted by procedural or
technological controls.
PRESERVATION STRATEGIES5
While electronic records enable information to be created, manipulated, disseminated,
located, and stored with increasing ease, preserving access to this information poses a
significant challenge. Unless preservation strategies are actively employed, this
information will rapidly become inaccessible. Choice of strategy will depend upon the
nature of the material and what aspects are to be retained.
Copying information without changing it, offers a short-term solution for preserving
access to digital material by ensuring that information is stored on newer media before
the old media deteriorates beyond the point where the information can be retrieved.
5
D.Bearman. Functional Requirements for Recordkeeping, in Electronic Evidence: Strategies for

Managing Records in Contemporary Organizations, Pittsburgh: Archives & Museum Informatics,
1994, http://www-prod.nla.gov.au/padi/topics/thesaurus.html.
21 CFR Part 11
150
The migration of digital information from one hardware/software configuration to

another or from one generation of computer technology to a later one, offers one method
of dealing with technological obsolescence.
While adherence to standards will assist in preserving access to digital information, it
must be recognized that technological standards themselves are evolving rapidly.
Technology emulation potentially offers substantial benefits in preserving the
functionality and integrity of digital objects. However, its practical benefits for this
application have not yet been well demonstrated.
Encapsulation, a technique for grouping together a digital object and anything else
necessary to provide access to that object, has been proposed by a number of researchers
as a useful strategy in conjunction with other digital preservation methods.
The importance of documentation as a tool to assist in preserving digital material is
universally agreed upon. In addition to the metadata necessary for resource discovery,
other sorts of metadata, including preservation metadata, describing the software,
hardware, and management requirements of the digital material, will provide essential
information for preservation.
The requirement for keeping every version of software and hardware, operating
systems, and manuals, as well as the retention of personnel with the relevant technology
skills, generally makes the preservation of obsolete technologies unfeasible.
Various frameworks designed to assist in managing the preservation of digital material
have been developed. These include tools designed to assist in the development of digital
preservation strategies. These will often entail the identification of the various stages in
the operation and future planning of a computer system when the provision of long-term
access should be considered.
ELECTRONIC RECORDS AUTHENTICITY5

The authenticity of electronic records refers to the degree of confidence that users can
have that the records are the same as those expected, based on a prior reference or
understanding of what they purport to be.
The digital environment poses particular challenges for establishing authenticity. This
is due to the ease with which electronic records may be altered and copied, which can
result in the possibility of a multiplicity of versions of a particular document.
Methods used in converting, storing, transmitting, or rendering electronic records may
result in distortions and therefore need to be documented. The process of migrating
information from one system or format to another may result in changes, which also need
to be recorded.
Aspects such as a documents functionality, its dependence on a particular software
application, and its relationship to other documents are all features that need to be
considered in the establishment of its authenticity.
A range of strategies for asserting the authenticity of digital resources have been
developed, and the choice of a particular method will depend upon the purpose for which
authenticity verification is required. Among these methods are the registration of unique
document identifiers and the inclusion of metadata within well-defined metadata
structures. Hashing and digital time stamping are public methods, authenticating the
Electronic records
151
existence of a document relating to a specific moment in time. Another class of methods

for establishing authenticity includes encapsulation techniques and encryption strategies.
A digital watermark can only be detected by appropriate software, and is primarily used
for protection against unauthorized copying. Digital signatures are used to record
authorship and identify the people who have played a role in a document.
STORAGE5
The choice of strategies and methods for storing digital information will determine its
future accessibility and usability. A comprehensive enterprise-wide data warehouse is the
subject-oriented, integrated, time-variant, nonvolatile collection of data, which can serve
as the central point of data integration for business intelligence and supports strategic,
tactical, and operational decision-making processes across the entire enterprise.
Options for storing digital information include the choice of media on which the
information will be stored, and decisions regarding the formats in which it will be stored.
The physical medium on which the information is stored should, ideally, be stable:
deterioration can be slowed by storing the digital information under stable environmental
conditions. General information regarding the preservation of access to information on
physical format digital materials is available on the National Library of Australias
Preserving
Access
to
Digital
Information
(PADI)
site
(http://wwwprod.nla.gov.au/padi/topics/thesaurus.html), as well as along with more specific
information about preserving access to material stored on optical disks and magnetic
media.
Strategies such as storing identical material in multiple locations and regularly backing
up will provide some protection against loss due to media failure, physical damage to
facility/ property, or human error.
While the choice of standard storage media can ameliorate the effects of technological
obsolescence, inevitably, old storage media will be superseded and migration, or some
other preservation strategy will be necessary to preserve access to the material. Similarly,
the selection of less volatile standard formats will assist in maintaining access and will
facilitate later migration; however, it must be remembered that in the information
technology environment, the standards themselves will change.
Whether digital information is stored on-line, near-line, or off-line will usually depend
upon the expected retrieval requirements. Little used off-line material may be stored on
magnetic tape, while the enhanced accessibility of disk storage may be chosen for highdemand on-line material.
Digital information stored in compressed form requires less space and thus reduces
storage costs. However, when an image is compressed and then decompressed, the
decompressed image is usually not quite the same as the original scanned image; this is
called lossy compression. Distortions can be particularly severe for high-compression
ratios; although the degree of loss can usually be reduced by adjusting the compression
parameters. The use of lossy compression for the long-term storage of digital material is
not recommended, due to the potential for irretrievable loss of information on
decompression, or on the migration of the digital material from one lossy compression
system to another.
Chapter 23
Electronic Signatures
REGULATORY GUIDELINE1
Electronic signature means a computer data compilation of any symbol or series of
symbols executed, adopted, or authorized by an individual to be the legally binding
equivalent of the individuals handwritten signature. 21 CFR Part 11.3(7)
GENERAL CONCEPTS
The application of an electronic signatures refers to the act of affixing, by electronic
means, a signature to an electronic record. Part 11 references two types of nonbiometricsbased electronic signatures: password/userid combination-based signatures and digital
signatures. In addition to these methods, authentication systems (e.g., passwords,
biometrics, physical feature authentication, behavioral actions, and token-based
authentication) can be combined with cryptographic techniques to form an electronic
signature. Refer to Chapter 15 for additional information on authentication.
Whichever method is used, there are some basic elements to be considered.

1. An electronic signature solution must make electronic signatures secure through the
use of a copy protection mechanism that makes it impossible to copy, cut, or paste
signatures and audit trails from an approved record. This is an element that is
necessary in order to ensure the integrity of digitally signed records.
1
FDA, Electronic Records; Electronic Signatures Final Rule, 62 Federal Register 13430 (March
20, 1997).
21 CFR Part 11
154
2. In the electronic environment, an electronic signature on an electronic record must

carry the same legal weight as an original signature on a paper-based document.
3. The electronic signature process involves:
authentication of the signer
a signature process that complies with the system design and software instructions
specified
the binding of the electronic signature to the electronic record
nonalterability after the signature has been affixed to the electronic record
4. The controls applicable to electronic signatures include:
uniqueness of the signature
signature/record linking
electronic signature security
password management (e.g., assignment, removal, loss management, aging)
in the case of dynamic passwords, it is required that the tokens that generate the
password are periodically tested
5. Electronic signature manifestation
The electronic signature must be displayed in human readable form, including
printouts and video displays:
immediately after the signature is executed
after displaying a signed record
when printing signed electronic record
The printed name of the signer, the date and time of the signing, and the meaning
associated with the signing must be displayed.
6. Multisigning
When an individual executes one or more signings that are not performed during
a single, continuous period of controlled system access, each signing must be
executed according to the following:
first signing: using both the userID and the password components
second and subsequent signings during a period of a continuous, controlled access:
either re-entry of the password alone or using both the userid and the password
components
One signature can apply to multiple data entries on a screen as long as the items that
the signature applies to are clearly indicated
PASSWORD-BASED SIGNATURES
In combination with PINs, Part 11.300 allows the use of password-based signatures. As
stated in Chapter 15, there are two PIN/password based authentication schemes: static
passwords and dynamic passwords. The same PIN/password combination used for
authentication may also be used for an electronic signature. The affixing of a signature to
155
a record should be an affirmative act that is deliberate, unique, and independent of the
authentication process, and that serves the ceremonial and approval functions of a
signature and establishes the sense of having legally consummated the transaction.
The records/signatures linking using password/PIN-based signatures is either centered
on the use of software locks, the storage of the electronic signature in a database table
separate from its associated record, or the storage of the signature within the subject
electronic record.
DIGITAL SIGNATURES2
Digital signatures are a form of electronic signatures suggested for open systems. A
digital signature provides the mechanism for verifying the integrity of the
signature/record linkage, and the identity of the signatory. The signature/record linkage is
a fundamental requirement for conformance with 21 CFR Part 11.70. Digital signatures
can be implemented in software, firmware, hardware, or any combination of these items.
Table 231 depicts the key features of
Table 231. Electronic Signature.

Requirement
Implementation
Digital signature
Ability to add attributes
If a digital signature is employed, three Continuity of signature capability
features must be implemented: Message Counter signatures
integrity, nonrepudiation, and user
Independent verifiability
authentication. Other implementation Interoperability
features are optional.
Message integrity
Multiple signatures
Nonrepudiation
Transportability
User authentication
digital signatures. Various technologies may fulfill one or more of the requirements
specified in Table 231. A comprehensive implementation requires the assurance of the
following:
records integrity
nonrepudiation
user authentication
2
O. Lpez, Overview of Technological Controls Supporting Security Requirements in Part 11,

presented at the PDA Annual Meeting, Washington, D.C., December 2001.
21 CFR Part 11
156
The PKCS3 describe how to digitally sign a message/record in such way that the
recipient can verify who signed it, and verify that it has not been modified since it was
signed. Refer to Figure 232. In summary:
1. The senders digital signature is associated with a pair of keys:
Private key
Public key
2. In order to sign a record, the record and the private key form the inputs to a hashing
process.
3. The output from the hashing process is a string bit pattern (message digest) that is
appended to the record. The plain text, the digital signature, and the senders digital
signing certificates4 are sent to the recipient.
4. At the recipient site, after the senders certificate is received, the CA digital signature
is checked to ensure that it was issued by someone that the recipient trusts.
5. The recipient of the transmitted record decrypts the message digest with the
originators public key, applies the same message hash function to the record, and then
compares the resulting message digest with the transmitted version.
6. Any modification to the record after it was signed will cause the signature verification
to fail (integrity).
7. If the digital signature was computed with a private key other than the one
corresponding to the public key used for verification, then the verification will fail
(authentication).
Figure 23-1. Digital Signature

Process.
3
Public-Key Cryptography Standards (PKCS) are a family of standards for public-key encryption
developed by RSA Laboratories. It describes the syntax for a number of data structures used with
public-key cryptography.
4
A signing certificate contains the public signing key assigned to an individual.
157
Figure 23-2. Signature Verification

Process.
In digital signatures, the private key signs, and the public key verify the authenticity of
signatures. For confidentiality, the public key encrypts messages, and the private key
decrypts messages.
Digital signatures are technologies that fully support the trustworthiness of signed
records. Some products on the market that support the integrity and signature
authentication to documents written in Microsoft Word, Microsoft Excel, and
Microsoft Outlook, Adobe Acrobat, JetForm FormFlow, PureEdge, XML, and HTML.
Some digital signature standards are:
Rivest-Shamir-Adleman (RSA) as specified in ANSI X9.31 Part either 1 (ISO 9796) or
PKCS #1
Digital Signature Standard (DSS) as specified in ANSI X9.30 Part 1 and NIST FIPS
PUB 186-2
Elliptic Curve DSA (ECDSA) as specified in ANSI X9.62
Chapter 24
Technologies Supporting Part 111,2
PAPER-BASED VERSUS ELECTRONIC-BASED SOLUTIONS
The record requirements discussed in Chapter 22 apply to both hard-copy records and to
electronic records. The privacy, authenticity, reliability, and nonrepudiation mechanism
used in a traditional paper-based solution consist of:
Condition
Solution
Privacy
Envelopes
Authenticity Notaries, strong ID, physical presence
Reliability
Signatures, watermarks, barcodes
Nonrepudiation Signatures, receipts, confirmations
The privacy, authenticity, reliability, and nonrepudiation mechanisms used in a digitalbased solution consist of:
Condition Solution
Privacy
Data encryption
Authenticity Digital signatures, digital certificates
Reliability
Hash algorithms, message digests, digital signatures
Nonrepudiation Digital signatures, audit trails
Electronic-based solutions can be used to support the security requirements contained in

Part 11. The following sections in this chapter provide an overview of the main
electronic-based solutions necessary to achieve trustworthy records, and to securing
computer resources to manage these records.
As will be noted, these solutions do not support the technological controls necessary to
implement audit trails.
1
O.Lpez, Technologies Supporting Security Requirements in 21 CFR Part 11, Part I,

Pharmaceutical Technology, February 2002.
2
O.Lpez, Technologies Supporting Security Requirements in 21 CFR Part 11, Part II,
Pharmaceutical Technology, March 2002.
Technologies supporting Part 11
159
HASH ALGORITHMS
Hashing refers to the process of computing a condensed message or record of any length
into a fixed length string using a one-way mathematical function so that one cannot
retrieve the message from the hash. The output of hashing process is called a message
digest. The probability of two different records generating the same message digest is 1
in 1087. Consequently, a message digest is unique and has a low probability of collisions.
Any change in a message, however minor, will result in a change in the message digest.
Regulatory guidance
Persons who use open systems to create, modify, maintain, or transmit electronic
records shall employ pro-cedures and controls designed to ensure the authenticity,
integrity, and, as appropriate, the confidentiality of electronic records from the point
of their creation to the point of their receipt. Such procedures and controls shall
include those identified in Sec. 11.10, as appropriate, and additional measures such
as document encryption and use of appropriate digital signature standards to ensure,
as necessary under the circumstances, record authenticity, integrity, and
confidentiality.
21 CFR Part 11.30
Because hashing is a one-way function and the output of the function has a low
probability of collisions, hashing can be used with the cryptographic product or services
family for authentication, nonrepudiation, and data integrity. An example of these is the
Digital Notary.3 Hashing is also a key element in the DSA.
Well-known hashing algorithms are:
MD2 and MD5. Both of these algorithms create a 128-bit message digest (RSA
Laboratories, RFCs 1319 and 1321, respectively)
SHA-1 (secure hashing algorithm) is a NIST-sponsored hashing function that has been
adopted by the U.S. government as a standard
Ripe-MD-160 is an algorithm from the European Community. It produces a 160-bit
message digest
3
Analogous to a public notary, Digital Notary Services provide trusted date-and-time stamp for a
document, so that someone can later prove that the document existed at a point in time. May also
verify the signature(s) on a signed document before applying the stamp.
21 CFR Part 11
160
DATA ENCRYPTION
Encryption refers to the process of scrambling input messages/records, called the
plaintext, with a user-specified password (password-based encryption algorithm) or key
(secret-key algorithm) to generate an encrypted output, called ciphertext. It is not possible
to recover the original plaintext from the ciphertext without the user-specified password
or key The algorithms comprising the user-specified password or key, and the plaintext,
are called ciphers. Encryption is mostly used to protect the privacy of messages or
records.
In the 1960s, Horst Feistel designed one of the first modern encryption algorithms at
IBM. Until recently all encryption algorithms were based on encrypting and decrypting
using the same private key. Only the owner knows the private key and will only share
this key with the parties he/she wants to communicate with. The sharing of a secret key
introduces an element of risk because compromising the secrecy of the private key may
also compromise the integrity of the data.
In 1976, Whitfield Diffie developed public-key encryption as an alternative to privatekey encryption. Public-key encryption is based on two halves of the same key that are
generated with special software at the same time. The key pair are mathematically related
so that the private key cannot be determined from the public key. Only one of the two
halves of the key pair is required to encrypt a message, with the other half being used for
decryption. In public key cryptography, one half of the key pair (the private key) is
assigned to an individual, and is closely guarded and securely stored on the users local
disk in either an encrypted format or as part of a token that interfaces with the computer.
The other half of the key is published in a public directory where all users can access it
and this therefore referred to as the public key.
Public-key cryptography, when properly implemented and used, enables people to
communicate in secrecy, and to sign documents, with almost absolute security and
without ever having to exchange a private key.
Provided there is a strong linkage between the owner and the owners public key, the
identity of the originator of a message or record can be traced to the owner of the private
key. Public-key encryption can play an important role in providing the necessary security
services including confidentiality, authentication, integrity of records, and digital
signatures.
PKCS is a family of standards for public-key encryption developed by RSA
Laboratories.4 They describe the syntax for several data structures used with public-key
cryptography.
One well-known product based on PKCS is the PKI. PKI is a combination of software,
encryption technologies, server platforms, workstations, policies, and services used for
administering certificates and public-private key pairs.
4
RSA Laboratories, in collaboration with Apple, Digital, Lotus, Microsoft, MIT, Northern
Telecom, Novell and Sun, developed a family of standards describing data structures used with
public key cryptography.
161
It enables organizations to protect the security of their communications and business

transactions when using networks. PKI is used to secure e-mails, Web browsers, VPNs,
and end applications.
In a traditional PKI architecture, a CA is a trusted party that vouches for the
authenticity for the entity in question. The CA notarizes public keys by digitally signing
the certificates using the CAs private key, which is linked to the entities5 concerned.
A Certificate Server is a repository for digital certificates. End applications that are
PKI-enabled verify the validity and access privileges of a certificate by checking the
certificates profile status, which is protected in the repository.
The Security Server provides services for managing users, digital certificate security
policies, and trust relationships in a PKI environment.
PKI architectures can be classified as one of three configurations: a single CA, a
hierarchy of CAs, or a mesh of CAs. Each of the configurations is determined by the
fundamental attributes of the PKI: the number of CAs in the PKI, where users of the PKI
place their trust, and the trust relationships between CAs within a multi-CA PKI.
Digital certificates are digitally signed data structures that contain information such as
an entitys name, public key, signature algorithm, and extensions. This information
resides in the Active Directory located on the Certificate Server. The International
Telecommunications Union X.509 standard is the most widely used digital certificate
specification. A sample X.509v3 compliant digital certificate6 data structure can be found
in Sidebar 241. The extensions can be used to tailor digital certificates to meet the needs
of the end applications.
End applications either have to be PKI-enabled, PKI-aware out of the box, or they
have to be enabled separately. The enabling process may involve using PKI-vendor
plug-ins (e.g., Entrust Technologies and Shym Technologies have plug-ins for SAP)
that can be added into the end application, or it may involve far more detailed
programming.
5
An entity can be a person, server, organization, account, or site.

X.509 Digital Certificate is the International Telecommunications UnionTelecommunication
Standardization Section (ITU-T) recommendation that defines a framework for the provision of
authentication services under a central control paradigm represented by a Directory. The
recommendation describes two levels: simple authentication, using a password as verification of
claimed identity, and strong authentication, involving credentials formed by using cryptographic
techniques, the certificate. The format of the certificate structure is defined along with
responsibilities of the CA with regard to establishing and maintaining trust.
6
21 CFR Part 11
162
Sidebar 241. Sample x.509.

With respect to component-level PKI interoperability, the contract developer/integrator
must understand that enabling an end application to operate with one vendors PKI
products does not necessarily ensure that the end application will also operate with a
different vendors PKI products. If companies do not want to find themselves stranded on
their own PK-island, then they must plan to integrate with other installations. However, if
a company makes their installation capable of accepting x.509v3 compliant digital
certificates, this does provide interoperability. The end application can accept such
certificates from multiple vendors CAs, provided that the certificates honour a consistent
certificate profile for their extension fields. A sample application-programming interface
to
a
PKI
service
can
be
found
at
the
following
address:
http://csrc.nist.gov/pki/pkiapi/welcome.htm.
The issue of PKI interoperability becomes complicated when it is compared with
interdomains. Interdomain interoperability involves several technologies and policyrelated challenges. Explaining these challenges is, however, beyond the scope of this
book.
Some encryption schemes include:
Rivest, Shamir, and Aldeman (RSA), 1977.
Diffie-Hellman
ElGamal Public Key system
Digital Signature Standard (DSS) that uses the DSA
RC4 used by Microsoft Kerberos (128-bit key length)
The FIBS approved algorithm includes:
DES
Triple DES, as specified in ANSI X9.52, Triple Data Encryption Algorithm Modes of
Operation
Skipjack
163
Table 24-1. Encryption Strength.

Key Length (bits) Amount of Time to Break
30
N/A. Can be cracked by brute force guessed
using a powerful PC
40
4 hours
56
22 hours (1999 figure)
64
Are probably breakable by powerful computers
80
Less than 7 months. Read the note for 512-bit key length
128
Less than 7 months. Read the note for 512-bit key length
128 AES
149 trillion years. Read the note on the new
encryption standard below
512
7 months (1999 figure, refer to the Internet address
http://www.cwi.nl/~kik/persb-UK.html)
1024
28,000 billion years (1996 figure. Using search techniques,
with a 100 MIPS (Million Instructions Per Second)
computerequivalent
to a 200-MHz Pentium)
The key length, in bits, determines the encryption strength of the cryptographic
algorithm. Refer to Table 241. For example, DES, which was adopted in 1977, uses a
56-bit key length and can be cracked by specialized computers in only a few hours. In
addition, the use of DES is not recommended due to the susceptibility to cryptographic
exhaustion attack. One replacement for DES, is Triple DES which uses two 56-bit keys.
Another example of the lack of strength of the cryptographic algorithms is related to
the plain text. A 40-bit encryption is the default for browsers and this can be easily
removed using an old desktop PC in four hours.
The U.S. government will be upgrading its data-encryption standard. The new AES,
which was selected after a four-year study, supports key sizes of 128, 192, and 256 bits.
According to NIST, if a theoretical machine were built fast enough to crack a standard
DES in one second, it would still take the same machine 149 trillion years to break a 128bit AES key.
Advances in the private sector include the XTR crypto-system. This system makes the
current schemes into more efficient and compact implementations.
The US government controls the export of cryptographic implementations. A recent
amendment to the Export Administration Regulations (15 CFR Parts 734, 740, 742, 770,
and 774) eases the restrictions on encrypted applications. The rules governing export can
be quite complex, since they consider multiple factors. In addition, encryption is a rapidly
changing field, and the rules may change from time to time. Questions concerning the
export of a particular implementation in the USA should be addressed to the appropriate
legal counsel.
DIGITAL SIGNATURES
21 CFR Part 11
164
WINDOWS OS
The Microsoft Windows 2000 is the first operating system to build PKI (adhering to
PKI standards) into its core, and allows the system programmers to establish and
maintain cryptographic-based security infrastructure and the foundation for a secure
network. Using Windows 2000 with other popular Microsoft applications allows users
to gain the following security capabilities:
secure e-mail using Outlook
secure Web access using Internet Explorer, Microsofts IIS, and Windows 2000
servers. The RSA algorithm is included as part of the Web browser by Microsoft
file encryption using Windows 2000 EFS
smartcard based single sign-on
VPNs using the IPSec/VPN capabilities within Windows 2000
The primary components in Windows 2000 that support PKI environments are
Certificate Services, Active Directory, standards-based PKI-enabled applications, and
Exchange Key Management Services.
Certificate Services is a core operating system service, which allows businesses to act
as their own CA. Based on the organizations CA approval instructions, the CA can issue
and manage a digital certificate to represent its e-business identities. Windows 2000
supports multiple levels of a CA hierarchy and cross-certification as well as off-line and
on-line CA for maximum flexibility. A CA can issue digital certificates for purposes such
as digital signatures, secure e-mail, and the authentication of Web servers using SSL7 or
TLS.
PKCS 1, 7, and 10 are used by the Windows 2000 Certificate Services. For example,
PKCS 10 describes how to construct a certificate request message. After the Windows
2000 Certificate Services processes the request, the operating system will issue a
x.509v3-compliant digital certificate that accepts or rejects the request. PKCS 10 may be
used to implement access controls in networked and end applications.
Items such as file permissions, registry settings, password usage, user rights, and other
issues associated with Windows 2000 security have a direct impact on the Certificate
Services security.
The Active Directory provides information from authoritative sources about people
and resources (such as employees, partners, customers, servers, roles, directories,
information, digital certificates, and so forth). It serves as the internal and external
Certificate Distribution System. Active Directory is also the centralized management
interface for digital certificate issuance.
Windows 2000 introduces the concept of the enterprise certificate authority. This
feature is integrated with the Active Directory and provides other features such as SSL
client mapping and smartcard logon.
7
Secure Sockets Layer (SSL) is considered to be the industry-standard protocol for secure, Webbased communications. A recent version includes data encryption between the server and the
browser and its support client authentication.
165
Some of the standards based PKI-enabled application include: Internet Explorer,

Encrypting File System, IPSec, Outlook, and Outlook Express.
The KMS is the component of Microsoft Exchange that allows the archiving and
retrieval of keys used to encrypt e-mail. Both Exchange Servers 5.5 and 2000 integrate
with a Windows 2000 CA for the issuance of x.509v3-compliant digital certificates.
The National Security Agency (refer to internet address: www.nsa.gov) has developed
security configuration guidance for Windows 2000 in order to provide information on
the services that are available in the Microsoft Windows 2000 environment and how to
integrate these services into their network architecture.
Chapter 25
All Together
Following the ISO/IEC International Standard,1 primary SLC processes can be used to
guide the reader on how to put together all the elements discussed in this book. It is not
the intention of this chapter to develop a paradigm or model for the regulated industry.
Primary SLC processes depend on who initiates or performs the development,
operation, or maintenance of software products. These primary parties are the acquirer,
the supplier, the developer, the operator, and the maintainer of software products. The
primary SLC processes consist of acquisition process, supply process, development
process, operation process, and maintenance process.
The following defines each process and the activities in this book associated with the
process.
ACQUISITION PROCESS
Defines the activities of the acquirer, the organization that acquires a system,
software product, or service.
The activities in this book associated with the Acquisition Process include:
Initiation
Chapter 8
Acquisition Planning
Request for Proposal
Chapter 7
Requirements gathering, including technical constraints Attachment E
Requirements gathering and specification
Contract Preparation Update
This activity is beyond the scope of this book.
All together
167
Supplier Monitoring
Chapter 9
In-process audits
Code reviews, inspections, audits
1
ISO1207:1995, Information technologySoftware life cycle processes.
Acceptance and Completion

Appendix E
FAT
SAT
SUPPLY PROCESS
Defines the activities of the supplier, the organization that provides the system,
software product, or software service to the acquirer.
The activities in this book associated with the Supply Process include:
Initiation; Preparation of Response; Contract
These activities are beyond the scope of this book.
Planning
Chapter 8
Development Planning
Execution and Control
Develop product based on Development Process
Maintain product based on Maintenance Process
Review and Evaluation
Chapter 9
In-process audits
21 CFR Part 11
168
Appendix E
FAT
SAT
Delivery and Completion
Appendix E
Formal release for use
DEVELOPMENT PROCESS
Defines the activities of the developer, the organization that defines an develops the
software product.
The activities in this book associated with the Development Process include:
Process Implementation
Chapter 5
Computer Validation Management Cycle
Chapter 6
Computer Validation Program Organization
Chapter 7
Computer Systems Validation Process
Chapter 8
Development Planning
Chapter 11
SLC Documentation
Appendix E
Sample Development Activities
Requirements Analysis
Chapter 7
Conceptualization
Development
All together
169
Requirements gathering
Part 11, Chapter 20
Security
Operation
Safety
System Specification
Chapter 22
Electronic records storage
Design Analysis
Chapter 7
Development
Technical design
Coding and Testing
Chapter 7
Development
Program build
Chapter 9
Integration
Chapter 9
Program built testing
Appendix E
Conduct Software Module (Unit) Testing and Integration Testing
Hardware Integration and Program Build
System Testing
Appendix E
FAT
Chapter 9
In-process audits
21 CFR Part 11
170
Installation
Chapter 10
Hardware Installation Qualification
Software Installation Qualification
Acceptance Support
Appendix E
SAT
OPERATION PROCESS
Defines the activities of the operator, the organization that provides the service of
operating a computer system in its live environment for its users.
The activities in this book associated with the Operation Process include:
Process Implementation and User Support
Chapter 12 and Appendix F
Procedural controls applicable to operation
Chapter 13
Change Control
Appendix E
Computer Systems Incidents
Operational Testing
Appendix E
Support to OQs and PQs
Support computer system release activities
System Operation
Chapter 18
Business Continuity
Operational Management
Periodic reviews
All together
171
MAINTENANCE PROCESS
Defines the activities of the maintainer, the organization that provides the service of
managing modifications to the software product to keep it current and in operational
fitness. This process includes the migration and retirement of the software product.
The activities in this book associated with the Maintenance Process include:
Process Implementation
Chapter 8
Maintenance Project Planning
Chapter 13
Change Control
Problem and Modification Analysis; Modification Implementation;
Maintenance Review/Acceptance
The maintenance activities and work products should be consistent with the Development
Process above.
Migration and Software Retirement
Attachment E
Retirement
Chapter 22
Records Preservation Strategies
Chapter 26
The Future
A recent paper1 addressed the future of computer validation. The authors of this paper
anticipated how the current industry events and trends would affect computer validation.
Technology is one of many trends constantly affecting computer validation. One area
requiring special attention is the introduction of expert systems into the FDA-regulated
environment. The simplest form of artificial intelligence generally used in applications
(such as mortgages, credit card authorization, fraud detection, e-commerce,
personalization) is the rule-based system, also known as the expert system (ES).
An expert system is a way of encoding a human expert's knowledge of a fairly narrow
area into a computer system. There are a couple of advantages to doing so. One is that the
human expert's knowledge becomes available to a large range of people. Another
advantage is that if you can capture the expertise of an expert in a field, then that
knowledge is not lost when he/she retires or leaves the firm.
Expert systems differ from standard procedural or object-oriented programs in that
there is no clear order in which code executes. Instead, the knowledge of the expert is
captured in a set of rules, each of which encodes a small piece of the experts knowledge.
Each rule has a left-hand side and a right-hand side.
The left-hand side contains information about certain facts and objects that must be
true in order for the rule to potentially be executed. Any rules whose left-hand sides
match in this manner at a given time are placed on an agenda.
One of the rules on the agenda is picked (there is no way to predict which one), and its
right-hand side is executed, and it is removed from the agenda. The agenda is then
updated (generally using a special algorithm called the Rete algorithm), and a new rule is
picked to execute. This continues until there are no more rules on the agenda.
Because the actions of the rule-based system tend to be hidden from view, people tend
not to realize just how extensively they are used.
A typical rule for a mortgage application might look something like this:
IF
(number-of-30-day-delinquencies >4)
AND (number-of-30-day-delinquencies <8)
THEN increase mortgage rate by 1%
As the reader can see, a rule bears a close resemblance to an if-then-else statement, but
unlike an if-then-else statement, it stands alone and does not fire in any predetermined
order relative to other if-then-else statements.
In a way, a rule-based system might almost be thought of as being similar to a
multithreaded system in that just as one doesnt know which thread will execute next,
1
R.Buihandojo, D.J.Bergeson, B.Bradley, P.DEramo, L.Huber, O.Lpez et al. The Future State of
Computer Validation, Pharmaceutical Technology, July/September 2001.
21 CFR Part 11
174
one doesn't know which rule will execute next. However, rule-based systems are
generally implemented as single-thread programs.
The advantage to this type of approach, as opposed to a procedural approach, is that if
the system is well designed, then the experts knowledge can be maintained fairly easily,
just by altering whichever rules need to be altered. Definitely, many rule-based systems
come along with a rules editor that allows for rules to be easily maintained by
nontechnical people.
Rules are generally implemented in something called a rules engine, which provides a
basic framework for writing rules and then for running them in the manner described
above.
In a standard procedural or object-oriented environment, sequencing is a key element
in the validation of a computer system. Sequencing encompasses instruction to operators,
operation/process sequencing, and Part 11-related operational checks. So, how are these
systems to be tested? Who will make sure that the calling sequence is correct?
Another area with immediate impact to computer validation is the current federal
legislation, Electronic Signatures in Global and National Commerce Act, on definitions
of electronic records. This definition is not consistent with Part 11. The current federal
legislation defines electronic records in more general terms than Part 11 does. The
primary element into consideration in the current federal legislation is whether the data is
accessible once put in storage, not considering the technology used. This definition will
consider data in transient memory as electronic records. One key element impacting
validation is the regulatory requirements to data stored in transient memory, including
audit trails.
Particularly for global companies, harmonization of SLC and development
methodology practices is another area that should be evaluated. To achieve this, a
common international standard may be used as a framework. The two (2) key
international standards are ISO900032 and ISO122073. Implementation of these
standards can be found in local standard bodies. The evolution of GAMP4 incorporates
these two standards. Refer to Figure 261. The framework of 12207 includes five
primary processes: acquired, supplying, developing, operating, and maintaining
software. It divides the five processes into activities, and the activities into tasks,
while placing requirements upon their execution. It also specifies eight supporting
processes: documentation, configuration management, quality assurance, verification,
validation, joint review, audit, and problem resolutionas well as four organizational
processesmanagement, infrastructure, improvement, and training. Now that
ISO9001:1994, ISO9002:1994 and ISO9003:1994 have been integrated into the new
ISO9001:2000, ISO90003 may be reviewed for consistency with the ISO standards.
Specifically for medical devices, the Medical Device Software Committee of the
Association for the Advancement of Medical Instrumentation (AAMI) has taken on the
task of developing a standard4 for use in the medical device business sector based upon
the framework established in ISO12207.
2
ISO90003:1997, Quality management and quality assurance standards Part 3, Guidelines for
the application of ISO9001:1994 to the development, supply, installation and maintenance of
computer software. Authors note: At the time of this book publishing, this guideline is in the
process of being updated to align with ISO9001:2000.
3
ISO12207:1995, Information technologySoftware life cycle processes.
The future
175
ANSI/AAMI SW68:2001, Medical device softwareSoftware life cycle processes.
Figure 261. ISO 9001 Evolution.

In the future, computer systems validation needs to be adaptable to new technologies
and regulation changes. In the future, the industry might be more experienced managing
computer technology suppliers and contract developers. The cost of implementing and
maintenance of computer systems can be streamlined by effectively leveraging the work
of computer technology suppliers and contract developers. The utilization of contract
developers verifications and testing to demonstrate that an automated solution meets the
defined specification is one of many activities to streamline the project costs.
Appendix A
Glossary of Terms
For additional terms, refer to the Glossary of Computerized System and Software
Development Terminology1 and A Globally Harmonized Glossary of Terms for
Communicating Computer Validation Key Practices.2
Abstraction This is a basic principle of software engineering, and enables an
understanding of an application and its design, and the management of complexity.
Acceptance Criteria The criteria that a system or component must satisfy to be accepted
by a user, customer, or other authorized entity (IEEE).
Acceptance Test Testing conducted to determine whether a system satisfies its
acceptance criteria and to enable the customer to determine whether to accept the
system (IEEE).
Access The ability or opportunity to gain knowledge of stored information (DOD
5015.2STD).
Application A group of computer instructions that is used to accomplish a business
function, control a process, or facilitate decision-making. Refer to Application
Software in Glossary of Computerized System and Software Development
Terminology, August 1995.
Approver(s) In the context of configuration management, the approver is the person(s)
responsible for evaluating the recommendations of the reviewers of deliverable
documentation, and for making a decision on whether to proceed with a proposed
change, and for initiating the change request implementation.
Auditor In the context of configuration management, the auditor is the person
responsible for reviewing the steps taken during a development or change
management process, to ensure that the appropriate procedures have been followed.
Audit Trail An electronic means of auditing the interactions between records within an
electronic system, so that any access to the system can be documented as it occurs, to
identify unauthorized actions in relation to the records, e.g., modification, deletion, or
addition (DOD 5015.2STD).
FDA, Glossary of Computerized System and Software Development Terminology, Division of

Field Investigations, Office of Regional Operations, Office of Regulatory Affairs, Food and Drug
Administration, August 1995.
2
Herr, Robert R. and Wyrick, Michael L., A Globally Harmonized Glossary of Terms for
Communicating Computer Validation Key Practices, PDA Journal of Pharmaceutical Science and
Technology, March/April 1999.
Appendix A
178
Authentication The process used to confirm the identity of a person, or to prove the
integrity of specific information. In the case of a message, authentication involves
determining the message source, and providing assurance that the message has not
been modified or replaced in transit (ABA).
Authenticity A condition that proves a record is authentic and/or genuine, based on its
mode (the method by which a record is communicated over space or time); on its form
(format and/or media that a record takes when received); state of transmission (the
primitiveness, completeness, and effectiveness of a record when initially set aside after
being made or received); and manner of preservation and custody (DOD 5015.2
STD).
Automated systems Include a broad range of systems including, but not limited to,
automated manufacturing equipment, automated laboratory equipment, process
control, manufacturing execution, clinical trials data management, and document
management systems. The automated system consists of the hardware, software, and
network components, together with the controlled functions and associated
documentation (GAMP).
Certification authority The authority (part of the public-key infrastructure) in a network
that issues and manages security credentials and public key for message encryption
and decryption from a certificate server (NARA).
Change Any variation or alteration in form, state, or quality. It includes additions,
deletions, or modifications impacting the hardware or software components used
affecting operational integrity, service level agreements, or the validated status of
applications on the system.
Cipher Series of transformations that converts plaintext to ciphertext using the Cipher
key.
Cipher key Secret cryptography key that is used by the Key Expansion routine to
generate a set of round keys.
Cipher text Data output from the Cipher or input to the Inverse Cipher.
Code audit An independent review of source code by a person, team, or tool to verify
compliance with software design documentation and programming standards.
Correctness and efficiency may also be evaluated (IEEE).
Code inspection A manual (formal) testing (error detection) technique where the
programmer reads source code, statement by statement, to a body that can ask
questions; for example, when analyzing the program logic, analyzing the code with
respect to a checklist of historically common programming errors, and analyzing its
compliance with coding standards. This technique can also be applied to other
software and configuration items (Myers/NBS).
Code review A meeting at which software code is presented to project personnel,
managers, users, customers, or other interested parties for comment or approval
(IEEE).
Code walkthrough A manual testing (error detection) technique where program (source
code) logic (structure) is traced manually (mentally) by a group or body with a small
set of test cases; the state of program variables is simultaneously manually monitored,
to analyze the programmers logic and assumptions (Myers/NBS).
Complexity In the context of this book, complexity means the degree to which a system
or component has a design or implementation that is difficult to understand and verify.
Appendix A
179
Sample factors for establishing the complexity of a system or a component can be

reviewed in Appendix D.
Compliance Compliance covers the adherence to application-related standards or
conventions or regulations in laws and similar prescriptions.
Configurable software Application software, sometimes general purpose, written for a
variety of industries or users in a manner that permits users to modify the program to
meet their individual needs (FDA).
Computer A functional unit that can perform substantial computations, including
numerous arithmetic operations and logical operations without human intervention.
Computer System (1) A system including the input of data, electronic processing, and
the output of information to be used either for reporting or automatic control (EU
PIC/S). (2) A functional unit, consisting of one or more computers and associated
peripheral input and output devices, and associated software, that uses common
storage for all or part of a program and also for all or part of the data necessary for the
execution of the program; executes user-written or user-designated programs;
performs user-designated data manipulation, including arithmetic operations and logic
operations; and execute programs that modify themselves during their execution. A
computer system may be a standalone unit or may consist of several interconnected
units (ANSI).
Computer Systems Validation (CSV) The formal assessment and reporting of quality
and performance measures for all the life-cycle stages of software and system
development, its implementation, qualification and acceptance, operation,
modification, requalification, maintenance, and retirement, such that the user has a
high level of confidence in the integrity of both the processes executed within the
controlling computer system(s), and in those processes controlled by and/or linked to
the computer system(s), within the prescribed operating environment(s) (MCA).
Concurrent Validation In some cases, a drug product or medical device may be
manufactured individually or on a one-time basis. The concept of prospective or
retrospective validation as it relates to those situations may have limited applicability.
The data obtained during the manufacturing and assembly process may be used in
conjunction with product testing, to demonstrate that the instant run yielded a finished
product meeting all of its specifications and quality characteristics (FDA).
Control System Included in this classification are Supervisory Control and Data
Acquisition Systems (SCADA), Distributed Control Systems (DCS), Statistical
Process Control systems (SPC), Programmable Logic Controllers (PLCs), intelligent
electronic devices, and computer systems that control manufacturing equipment or
receive data directly from manufacturing equipment PLCs.
Criticality In the context of this book, criticality means the regulatory impact to a system
or component. Sample factors that establish the criticality of a system or a component
can be reviewed in Appendix D.
Custom-built software Also known as a Bespoke System, Custom-Built Software is
software produced for a customer, specifically to order, to meet a defined set of user
requirements (GAMP).
Digital certificate A credential issued by a trusted authority. An entity can present a
digital certificate to prove its identity or its right to access information. It links a
public-key value to information that identifies the entity, associated with the use of the
Appendix A
180
corresponding private key. Certificates are authenticated, issued, and managed by a

trusted third party called a CA.
Documentation Manuals, written procedures or policies, records, or reports that provide
information concerning the use, maintenance, or validation of a process or system
involving either hardware or software. This material may be presented from electronic
media. Documents include, but are not limited to, Standard Operating Procedures
(SOPs), Technical Operating Procedures (TOPs), manuals, logs, system development
documents, test plans, scripts and results, plans, protocols, and reports. Refer to
Documentation and Documentation, level of in the Glossary of Computerized
System and Software Development Terminology, August 1995.
Emergency change A change to a validated system determined necessary to eliminate an
error condition that prevents the use of the system and interrupts business function.
Emulation The process of mimicking, in software, a piece of hardware or software so
that other processes think that the original equipment/ function is still available in its
original form. Emulation is essentially a way of preserving the functionality of, and
access to, digital information that might otherwise be lost due to technological
obsolescence.
Encryption The process of converting information into a code or cipher. A secret key, or
password, is required to decrypt (decode) the information, which would otherwise be
unreadable.
Entity A software or hardware product that can be individually qualified or validated.
Establish Establish is defined in this book as meaning to define, document, and
implement.
Factory Acceptance Test (FAT) An acceptance test in the suppliers factory, usually
involving the customer (IEEE).
Field Devices Hardware devices typically located in the field at or near the process,
necessary for bringing information to the computer or implementing a computerdriven control action. Devices include sensors, analytical instruments, transducers, and
valves.
GxP A global abbreviation intended to cover GMP, GCP, GLP, and other regulated
applications in context.
GxP Computerized Systems A computerized system that performs regulated operations
that are required to be formally controlled under the Federal Food, Drug, and
Cosmetic Act, the Public Health Service, and/or an applicable regulation.
Hybrid systems In the context of Part 11, hybrid computer systems save required data
from a regulated operation to magnetic media. However, the electronic records are not
electronically signed. In summary, in hybrid systems some portions of a record are
paper and some are electronic.
Impact of change The impact of change is the effect of change on the GxP computerized
system. The components by which the impact of change is evaluated may include, but
not be limited to, business considerations, resource requirements and availability,
application of appropriate regulatory agency requirements, and criticality of the
system.
Inspection A manual testing technique in which program documents [specifications
(requirements, design), source code or user's manuals] are examined in a very formal
and disciplined manner to discover any errors, violations of standards or other
Appendix A
181
problems. Checklists are typical vehicles used in accomplishing this process. Refer to
Chapter 9.
Installation qualification Establishing confidence that process equipment and ancillary
systems are capable of consistently operating within established limits and tolerances
(FDA).
Key practices Processes essential for computer validation that consists of tools,
workflow, and people (PDA).
Legacy systems Production computer systems that are operating on older computer
hardware or are based on older software applications. In some cases, the vendor may
no longer support the hardware or software.
Logically secure and controlled environment A computing environment, controlled by
policies, procedures, and technology, which deters direct or remote unauthorized
access that could damage computer components, production applications, and/or data.
Major change A change to a validated system that is determined by reviewers to require
the execution of extensive validation activities.
Metadata Data describing stored data; that is, data describing the structure, data
elements, interrelationships, and other characteristics of electronic records (DOD
5015.2STD).
Migration Periodic transfer of digital materials from one hardware/software
configuration to another, or from one generation of computer technology to a
subsequent generation.
Minor change A change to a validated system that is determined by reviewers to require
the execution of only targeted qualification and validation activities.
Model A model is an abstract representation of a given object.
Module testing Refer to Testing, Unit in the Glossary of Computerized System and
Software Development Terminology, August 1995.
NEMA enclosure Hardware enclosure (usually a cabinet) that provides different levels
of mechanical and environmental protection to the devices installed within it.
Noncustom Purchased Software Package A generally available, marketed software
product that performs specific data collection, manipulation, output, or archiving
functions. Refer to Configurable, off-the-shelf software in the Glossary of
Computerized System and Software Development Terminology, August 1995.
Operating environment All outside influences that interface with the computer system
(GAMP).
Ongoing evaluation A term used to describe the dynamic process employed after a
systems initial validation that can assist in maintaining a computer system in a
validated state.
Operational testing Refer to Operational Qualification in the Glossary of
Computerized System and Software Development Terminology, August 1995.
Operating system Software controlling the execution of programs and providing
services such as resource allocation, scheduling, input/output control, and data
management. Usually, operating systems are predominantly software, but partial or
complete hardware implementations are possible (ISO).
Password A character string used to authenticate an identity. Knowledge of the password
that is associated with a userID is considered proof of authorization to use the
capabilities associated with that userID (CSC-STD-00285).
Appendix A
182
Packaged software Software provided and maintained by a vendor/supplier that can

provide general business functionality or system services. Refer to Configurable, offthe-shelf software in the Glossary of Computerized System and Software
Development Terminology, August 1995.
Periodic review A documented assessment of the documentation, procedures, records,
and performance of a computer system to determine whether it is still in a validated
state and what actions, if any, are necessary to restore its validated state (PDA).
Personal Identification Number A PIN is an alphanumeric code or password used to
authenticate the identity of an individual.
Physical environment The physical environment of a computer system comprising the
physical location and the environmental parameters in which the system physically
functions.
Planned change An intentional change to a validated system for which an
implementation and evaluation program is predetermined.
Policy A directive that usually specifies what is to be accomplished.
Predicate regulations Federal Food, Drug, and Cosmetic Act, the Public Health Service
Act, or any FDA Regulation, with the exception of 21 CFR Part 11. Predicate
regulations address the research, production, and control of FDA regulated articles.
Procedural controls (1) Measures taken to ensure the trustworthiness of records and
signatures established through the implementation of standard operating procedures
(SOPs). (2) Written and approved procedures providing appropriate instructions for
each aspect of the development, operations, maintenance, and security applicable to
computer technologies. In the context of regulated operations, procedural controls
should have QA/QC controls equivalent to the applicable predicate regulations.
Process system The combination of the process equipment, support systems (such as
utilities), and procedures used to execute a process.
Production environment The operational environment in which the system is being
used for its intended purpose, i.e., not in a test or development environment.
Production verification (PV) Documented verification that the integrated system
performs as intended in its production environment. PV is the execution of selected
Performance Qualification (PQ) tests in the production environment using production
data.
Prospective validation Validation conducted prior to the distribution of either a new
product, or product made under a revised manufacturing process, where the revisions
may affect the products characteristics. (FDA)
Qualification (1) Action of proving that any equipment works correctly and actually
leads to the expected results. The word validation is sometimes widened to incorporate
the concept of qualification (EU PIC/S). (2) Qualification is the process of
demonstrating whether a computer system and associated controlled
process/operation, procedural controls, and documentation are capable of fulfilling
specified requirements.
Qualification protocol A prospective experimental plan stating how qualification will be
conducted, (including test parameters, product characteristics, production equipment,
etc.) and decision points on what constitutes an acceptable test. When executed, a
protocol is expected to produce documented evidence that a system or subsystem
performs as required.
Appendix A
183
Qualification reports These are test reports that evaluate the conduct and results of the
qualification carried out on a computer system.
Raw data The original data that has not been manipulated or data that cannot be easily
derived or recalculated from other information.
Record A record consists of information, regardless of medium, detailing the transaction
of business. Records include all books, papers, maps, photographs, machine-readable
materials, and other documentary materials, regardless of physical form or
characteristics, made or received by an Agency of the United States Government
under Federal law; or in connection with the transaction of public business and
preserved or appropriate for preservation by that Agency or its legitimate successor as
evidence of the organization, functions, policies, decisions, procedures, operations, or
other activities of the Government or because of the value of data in the record (44
U.S.C. 3301, reference (bb)).
Record owner Record owner means a person or organization who can determine the
contents and use of the data collected, stored, processed or disseminated by that party
regardless of whether or not the data was acquired from another owner or collected
directly from the provider.
Record reliability A reliable record provides contents that can be trusted as a full and
accurate representation of the transactions, activities, or facts to which they attest, and
can be depended upon in the course of subsequent transactions or activities (NARA).
Regulated operations Process/business operations carried out on an FDA-regulated
product covered in a predicated rule.
Replacement The implementation of a new Part 11-compliant system after the retirement
of an existing system.
Requalification Repetition of the qualification process or a specific portion thereof.
Remediate software hardware and/or procedural changes employed to bring a system
into compliance with 21 CFR Part 11.
Retirement phase The period in the SLC in which plans are made and executed to
decommission or remove a computer technology from operational use.
Retrospective evaluation Establishing documented evidence that a system does what it
purports to do, based on an analysis of historical information. The process of
evaluating a computer system currently in operation against standard validation
practices and procedures. The evaluation determines the reliability, accuracy, and
completeness of a system.
Retrospective validation See Retrospective evaluation.
Site Acceptance Test (SAT) An acceptance test at the customers site, usually involving
the customer (IEEE).
Software development standards Written policies or procedures that describe practices
a programmer or software developer should follow in creating, debugging, and
verifying software.
Source Code The human readable version of the list of instructions (programs) that
enable a computer to perform a task.
Specification A document that specifies, in a complete, precise, and verifiable manner,
the requirements, design, behaviour, or other characteristics of a system or component;
and often the procedures for determining whether such provisions have been satisfied
(IEEE).
Appendix A
184
Static analysis (1) Analysis of a program performed without executing the program
(NBS). (2) The process of evaluating a system or component based on its form,
structure, content, documentation (IEEE).
Standard instrument software Software driven by non-user-programmable firmware,
which is configurable (GAMP).
Standard software packages A complete and documented set of programs supplied to
several users for a generic application or function (ISO/IEC 238220:1990).
System (1) People, machines, and methods organized to accomplish a set of specific
functions (ANSI). (2) A composite entity, at any level of complexity, of personnel,
procedures, materials, tools, equipment, facilities, and software. The elements of this
composite entity are used together within the intended operational or support
environment, to perform a given task or achieve a specific purpose, support, or
mission requirement (DOD).
System backup The storage of data and programs on a separate media, and stored
separately from the originating system.
System Life Cycle (SLC) The period of time commencing from when the system
product is recommended, until the system is no longer available for use or is retired.
System owner The person(s) who have responsibility for the operational system, and
bear the ultimate responsibility for ensuring a positive outcome of any regulatory
inspection or quality audit of the system.
System retirement The removal of a system from operational usage. The system may be
replaced by another system or may be removed without being replaced.
System software See Operating System.
Technical controls Measures taken to ensure the trustworthiness of records and
signatures established through the application of computer technologies.
Test report Document presenting test results and other information relevant to a test
(ISO/IEC Guide 2:1991).
Test script A detailed set of instructions for execution of the test. This typically includes
the following:
Specific identification of the test
Prerequisites or dependencies
Test objective
Test steps or actions
Requirements or instructions for capturing data (e.g., screen prints, report printing)
Pass/fail criteria for the entire script
Instructions to follow in the event that a nonconformance is encountered
Test execution date
Person(s) executing the test
Review date
Person reviewing the test results
For each step of the test script, the item tested, the input to that step, and the expected
result, are indicated prior to execution of the test. The actual results obtained during these
steps of the test are recorded on (or attached to) the test script. Test scripts and results
may be managed through computer-based electronic tools. Refer to Test case in the
Glossary of Computerized System and Software Development Terminology, August 1995.
Appendix A
185
Test Nonconformance A test nonconformance occurs when the actual test result does
not equal the expected result or an unexpected event (such as a loss of power) is
encountered.
Training plan Documentation describing the training required for an individual based on
his or her job title or description.
Training record Documentation (electronic or paper) of the training received by an
individual that includes, but is not limited to, the individuals name or identifier; the
type of training received; the date the training occurred; the trainers name or
identifier; and an indication of the effectiveness of the training (if applicable).
Trustworthy Reliability, authenticity, integrity, and usability are the characteristics used
to describe trustworthy records from a record management perspective (NARA).
Transient memory Memory that must have a constant supply of power or the stored data
will be lost.
Unplanned (emergency) change An unanticipated necessary change to a validated
system requiring rapid implementation.
User back-up/alternative procedures Procedure describing steps to be taken for the
continued recording and control of the raw data in the event of a computer system
interruption or failure.
Unit A separately testable element specified in the design of a computer software
element. Synonymous to component, module (IEEE).
UserID A sequence of characters that is recognized by the computer and that uniquely
identifies one person. The UserID is the first form of identification. UserID is also
known as a PIN or identification code.
Validated The term validated is used to indicate a status that designates that a system
and/or software is compliant with all regulatory requirements.
Validation Action of proving, in accordance with the principles of Good Manufacturing
Practice, that any procedure, process, equipment, material, activity, or system actually
leads to the expected results (see also qualification) (EU PIC/S).
Validation coordinator A person or designee responsible for coordinating the validation
activities for a specific project or task.
Validation protocol A written plan stating how validation will be conducted, including
test parameters, product characteristics, production equipment, and decision points on
what constitutes acceptable test results (FDA).
Validation plan A multidisciplinary strategy from which each phase of a validation
process is planned, implemented, and documented to ensure that a facility, process,
equipment, or system does what it is designed to do. May also be known as a system
or software Quality Plan.
Validation Summary Report (VSR) Documents confirming that the entire project
planned activities have been completed. On acceptance of the Validation Summary
Report, the user releases the system for use, possibly with a requirement that
continuing monitoring should take place for a certain time (GAMP).
Verification The process of determining whether or not the products of a given phase of
the SLC fulfill the requirements established during the previous phase.
Work products The intended result of activities or processes (PDA).
Worst case A set of conditions encompassing upper and lower processing limits and
circumstances, including those within standard operating procedures, which pose the
Appendix A
186
greatest chance of process or product failure when compared to ideal conditions. Such
conditions do not necessarily induce product or process failure (FDA).
Appendix B
Abbreviations and Acronyms
ABA
American Bar Association
ANDAs
Abbreviated New Drug Applications
ANSI
American National Standard Institute
CA
Certification Authority
CAPA
Corrective and Preventive Actions
CFR
Code of Federal Regulation
cGMP
current Good Manufacturing Practices
CPG
FDA Compliance Policy Guide
DES
Data Encryption Standard
DRM
Device Master Record
DSA
Digital Signature Algorithm
EFS
Encrypting File System
EU
European Union
EU PIC/S
European Union Pharmaceutical Inspection Cooperation

Scheme
FAT
Factory Acceptance Test
FD&C Act
US Food Drug and Cosmetic Act
FDA
US Food and Drug Administration
FR
US Federal Register
GAMP
Good Automated Manufacturing Practices
GCP
Good Clinical Practices
GLP
Good Laboratory Practices
GMPs
United States Good Manufacturing Practices
GxP
A global abbreviation intended to cover GMP, GCP, GLP,

and other regulated applications. In context GxP can refer
to one specific set of practices or to any combination of the
three.
I/Os
Inputs and outputs
IEEE
Institute of Electrical & Electronic Engineers
IIS
Internet Information Services
ISO
International Organization for Standardization
IT
Information Technology
Appendix B
188
KMS
Key Management Service
LAN
Local Area Network
MCA
UK Medicine Control Agency
MES
Manufacturing Execution System
NARA
National Archives and Records Administration
NBS
National Bureau of Standards
NDAs
New Drug Applications
NEMA
National Electrical Manufacturers Association
NIST
National Institutes of Standards and Technology
OECD
Organization for Economic Cooperation and Development
OSHA
US Occupational Safety & Health Administration
PDA
Parenteral Drug Association
P&ID
Process and Instrumentation Drawings
PIC/S
EU Pharmaceutical Inspection Cooperation Scheme
PIN
Personal Identification Number
PKCS
Public Key Cryptography Standards
PKI
Public-Key Infrastructure
PLC
Programmable Logic Controller
R&D
Research and Development
RFP
Request for Proposal
SAP
Systems, Application and Products
SAS
The Statistical Analysis System licensed by the SAS

Institute, Inc.
SAT
Site Acceptance Test
SCADA
Supervisory Control and Data Acquisition
SLC
System Life Cycle
SOPs
Standard Operating Procedures
SQA
Software Quality Assurance
SQE
Software Quality Engineering
SSA
US Social Security Administration
SSL
Secure Sockets Layer
TLS
Transport Layer Security
UK
United Kingdom
UPS
Uninterruptable Power Supply
US
United States
VPN
Virtual Privates Network
Appendix C
Applicability of a Computer Validation
Model
This appendix explains how to utilize the Part 11 computer validation model (Figure 41)
and the procedures necessary to determine the model parameters and to apply the results.
The elements that need to be taken into account during the implementation of a
computer system are: (i) the type of system (open or closed), (ii) security functions, audit
trails, (iii) the operation controlled by the computer system, and (iv) the technology
necessary to support the operation and Part 11. These last two elements are not identical
for all systems.
An element beyond the scope of the model is the retention of electronic records, but
the model can be used to verify and validate the implementation of the system(s) that will
hold these records. A subset of the above requirements are applicable to hybrid computer
systems,1 and the implementation of these requirements have been discussed in other
articles.2,3 The fundamental requirements that Part 11 establishes are as follows:
1. Open/Closed Systems
2. Security
a. System security
c. UserID and password maintenance
UserID and password security
Password assignment
4. Audit Trails
a. Audit mechanism
b. Metadata
1
In hybrid systems, some portions of a record are paper and some electronic.
Lpez, O, Automated Process Control Systems Verification and Validation, Pharmaceutical
Technology, September 1997.
3
Lpez, O, Implementing Software Applications Compliant with 21 CFR Part 11, Pharmaceutical
Technology, March 2000.
2
Appendix C
190
a. Electronic signature without biometric/behavioral identification
b. Electronic signature with biometric/behavioral identification
When determining which requirements need to be implemented, verified, and tested,
the following points need consideration: whether the records are held in transient
memory; whether the system is a hybrid system; and whether the system manages
electronic signatures. The following sections describe the implementation of the
model.
SAMPLE SITUATIONS
The objective of the following sample situations is to demonstrate the applicability of the
new computer systems validation model.
First Case: Records Held in Transient Memory
Two examples of this case are (i) the temperature profile from an autoclave, and (ii)
standalone PLC-based systems. All I/Os are saved in the controllers transient memory.
Neither the inputs (e.g., the operator keyboard) nor the outputs (e.g., the operator display)
are retrieved from, or saved to, durable media. If the record has not been saved yet to
durable media, it will be vulnerable to accidental or intentional alteration.
Because no records are yet saved to a durable media, Part 11 is not applicable to these
computer systems, or to other computer systems with similar configuration. However,
Part 11 may be used as the model to validate the software, hardware, and any associated
interfaces. In this case, the following elements in Part 11 are applicable:
1. Operational checks
2. System security
3. UserID and password security
4. UserID and password maintenance
5. Password assignment
6. Location checks
7. Authority checks
8. Document control
10. Record retention/protection
Operational checks are application dependent. Refer to Chapter 20. The reader may
realize that the application-dependent requirements become part of the model through the
operational verifications process.
Appendix C
191
In this first case, system security is associated with preventing the accidental or
intentional alteration and corruption of the data to be displayed on the screen, or be used
to make a decision to control the operation. To avoid accidental or intentional loss of
data, the data collected must be defined, along with the procedures used to collect it, and
the means to verify its integrity, accuracy, reliability, and consistency. A failure modesand-effects analysis (FMEA) is one of many methods used to uncover and solve these
factors. For example, to avoid data corruption, an ongoing verification program (Chapter
18) should be implemented.
Logical access controls for computer systems, including authority checks, form
another security aspect. Authority checks are based on the various roles and
responsibilities assigned to individuals known to the system. Computer systems must be
designed to make distinctions between the controlled access (a) to the system, (b) to
functions in the system, and (c) to input and output devices used by the system.
Location checks, when applicable, enable the application software to determine
whether the input being generated by a particular device is appropriate.
Passwords are one of many methods used to authenticate authorized users. As security
of computer systems performing regulated operations is so very critical, expectations for
activities (such as assignment, security, and maintenance of passwords) are clearly
established by the FDA. Before Part 11, the FDA expected4 that passwords would be:
Periodically changed, not be reassigned and not reused
Unrecognizable as reflections of the users personal life
Not shared by other users
The FDA expects written procedures or company policies to be in place covering the use
and security of passwords, supported by training on the subject. To ensure that computer
systems are secure, password files may be encrypted or otherwise secured so that the
password cannot be read by ordinary means. Password file encryption is commonly used
in modern operating systems.
In general, the assignment of passwords and their maintenance is a combination of
written procedures and the technologies used to support them. Sample procedures include
password aging, minimum password length, password uniqueness, default password
management, and account lockout after a reasonable number of unsuccessful log-in
attempts.
The focus of the validation program is usually the quality attributes of the system
implementation, change control, and the verification and testing of modifications made to
the baseline system. The configuration management of the system, including its
documentation, is a key area of concern. Specifically, documentation management is
extremely critical for the information contained in both the master production records,
and in the application.
4
FDA, draft of Guideline for the Validation of Blood Establishment Computer Systems, September
28, 1993.
Appendix C
192
System documentation comprises records relating to system operation and

maintenance, from the high-level design documents to the end-user manuals. The control
of system documentation includes an assessment of the documentation that may be
affected when a modification to the computer system is suggested. The high-level design
documentation provides the input for maintenance; user manuals provide the instructions
on how to operate the system.
According to Part 11, a computer system is considered closed if the person responsible
for the content of electronic records only allows their access by authorized parties. In
addition, electronic records must be monitored through the computer systems software
with its required logon, security procedures, and audit trails. For PLC-based systems, it is
very common to communicate with remote systems via the telephone system, accessed
by modem. A connection such as this can and may be used by the system
developer/integrator to perform diagnostics: this practice must be disallowed due to its
inherent security and configuration management weaknesses. If this type of access is
allowed, the modem must have strict access controls and restrictions, including protective
software, and access restrictions on such systems be designed to prevent unauthorized
modifications.
In this example, the user would record on paper each significant step in the operation.
The retention and protection of the paper-based record(s) is based on the retention
requirements that can be found in the applicable predicate rule.
Second Case: Hybrid Computer Systems
Hybrid computer systems save data required by the existing regulations onto durable
media. However, these electronic records are not electronically signed. An example of
this configuration is provided by computer systems generating paper-based reports,
which are printed and manually signed.
For hybrid systems, the electronic record requirements in Part 11 (Sub-Part B) are
applicable. The electronic records are maintained and retained in electronic form for the
period established by the predicate regulation that require these records. The elements of
Part 11 applicable to hybrid systems are:
1. Audit trails and metadata
2. System security
3. UserID and password security
4. UserID and password maintenance
5. Password assignment
6. Operational checks
7. Authority checks
8. Location checks
9. Document control
11. Record retention/protection
12. Signature/record linkage
Appendix C
193
A key element of such a system is the completeness of the recorded information. As part
of the validation process, the completeness of computer-generated electronic records is
verified.
An audit trail is a journal recording any modifications carried out by the users, or by
the processes operating on behalf of the user, to the electronic records. This mechanism
provides the capability to reconstruct modified data and must not obscure the previously
entered data. The tracking mechanism includes adding computer-generated date and time
stamps to the record, to indicate when the record was modified, the types of
modifications performed, and the identity of the person performing the modification to
the record.
The implicit requirements in Part 11 include the availability of the tools necessary to
display and/or print audit trail records, and associated metadata, in human readable form.
Another important use of audit trails is the protection from subsequent unauthorized
alteration and destruction, and to permit the detection of security violations and after-thefact investigations into their cause.
The point at which data becomes an electronic record should reflect its origin and
intended purpose. For manufacturing systems, this may be the initial point of data
acquisition and this point would also form the basis for the recording of audit trails for
changes to a record.
Third parties can provide custom code, which can be integrated within most common
database systems in order to implement audit trails. This approach provides a
technological foundation for the proper administration of electronic records, without
replacing the application that manages them.
For example, audit trails may be implemented by using database triggers. A trigger is
a feature of the Oracle database, which allows code written by the user to be executed
based on an event. The concept implies that the edition of a record will produce an event.
The event may be defined as an insert, update, or deletion of data from a database table.
An alternative method is to replace the complete application with an out-of-the-box
application, which includes an audit-trail feature.
Another alternative is to prevent modifications to the records at the data collection and
acquisition domains. The permanent repository system may be located at the business
logistic systems level.5 At this level, the electronic records are either maintained while in
active use, or are maintained off-line when their use is less frequent. This setting must be
evaluated and, preferably, documented as part of a company wide strategy. The selected
settings require procedural and technological controls.
Data is most often transferred using available drivers. The transfer/transcription
process, from the secondary repository to the permanent repository, shall be completely
qualified. The qualification must include all of the appropriate documentation and
supporting data. The objective of the qualification must be to ensure the accurate transfer
of electronic records and their associated content, context, and structure.
5
Enterprise/Control Integration standard, ISA SP95.01. The plant management level is the plan
repository area (e.g., database server(s)) that unifies production, quality control, inventory, and
warehouse data.
Appendix C
194
An important item to be addressed during the implementation of every computer

system is the long-term archiving of data and any associated metadata. The hardware and
software supporting the repository must be robust, but also flexible enough to withstand
the regulatory electronic records retention requirements.
The use of audit trails is not restricted to the modifications of the records. Audit trails
may also be used to record operator entries and actions during the operation of the
computer system, which can be recorded without the application of electronic signatures.
In addition to security elements referenced in the previous case, additional security
functions are needed to cover electronic records. External safeguards must be in place to
ensure that access to the computer system and associated data is restricted to authorized
personnel.
Access to records at the primary and secondary repositories must be restricted and
monitored through the systems software with its required logon, security procedures, and
audit trail. If remote access is provided via external software, the applications must enter
through the same protective security software as that used for local access.
In this case, reports are paper-based. The operator/user may record each significant
step of the computer-related operation and/or critical outputs. Paper records and
associated electronic records are retained, based the requirements of the applicable
predicate rule. In hybrid systems, one approach to the signature/records linkage is by
referencing the file(s) as part of the report. The reference may include the name, and
creation date and time of the files associated with the batch report. Modifications to any
referenced electronic file will invalidate the approval of the associated batch report.
It is very important to ensure that having a signed paper-based record (e.g., a report)
associated with one or more electronic records, the electronic file(s) associated with these
records must not be deleted. The Part 11 regulation is very specific regarding this issue.
Electronic files in hybrid systems must be maintained electronically. The interpretation of
this requirement in the medical devices CGMP regulations (21 CFR 820) is different. Part
820 requires that the results of acceptance activities are recorded, but are not necessarily
raw data, and these results must have audit trails. This interpretation is contained in the
medical device quality system preamble (pp. 52631 and 52646).
Third Case: Implementation of Electronic Signatures
The application of an electronic signature refers to the act of affixing, by electronic
means, a signature to an electronic record. Part 11 refers to two types of nonbiometricsbased electronic signatures: password/ID code-based signatures, and digital signatures. In
addition to these methods, authentication systems (e.g., passwords, biometrics, physical
feature authentication, behavioral actions, and token-based authentication) can be
combined with cryptographic techniques to form an electronic signature.
Signed records must first conform to the requirements established in Part 11 Sub-Part
B. In addition, verification and/or testing is required on the following:
2. Electronic signature security
3. Electronic signatures without biometric/behavioral identification
Appendix C
195
4. Electronic signatures with biometric/behavioral

5. Operational checks identification
6. Signature manifestation
7. Signature purpose
8. Signature binding and linkage
A certification must be sent to the FDA by the company, which informs the FDA that
electronic signatures will be used. This certification can be a global statement of intent,
meaning that a single company certification can cover all systems, all applications, all of
the electronic signatures, for all employees in a company, for all the firms locations
anywhere in the world. The Regulatory Affairs Unit of the company will usually send
this certification to the FDA.
Systems that use a password/ID code-based nonbiometrics electronic signature are
common, and use both a PIN and a password as authentication and as an electronic
signature. The combination of the PIN and password must be unique. Generally,
companies and organizations ensure that the PIN is unique so that if, by coincidence, two
persons create the same password, this does not result in two identical electronic
signatures. Furthermore, firms generally establish unique (but not confidential) PINs.
Biometrics is the use of physical characteristics such as a voiceprint, fingerprint, hand
geometry, and iris or retina pattern, as a means of uniquely identifying an individual.
When biometrics measures are applied in combination with other controls (such as access
cards or passwords), the reliability of authentication controls takes a giant step forward.
The use of biometrics/behavioral identification is less common in the FDA-regulated
industry. This access control technology does not change the authentication of users; it
changes the implementation of how to authenticate users.
Part 11 requires a signature manifestation to be executed as part of printing and
displaying signed records. The regulation requires that the name of the signer, the date
and time of the executed signature, and the purpose associated with the signing, be
displayed/printed in human readable forms.
As part of the action of signing an electronic record, the purpose of the signature must
be identified and be an element of each signed record. In the drug CGMP arena, there are
relatively few purposes for a signature and these include authored by, reviewed by,
and approved by The execution of an action such as a production step does not require
an electronic signature, and can be documented via the audit trail.
Special attention must be paid to the signature binding requirement of Part 11. The
current requirement states that electronic signatures must not be replaced or removed.
This requirement ensures the integrity of the signed records. In addition to the access
control technologies and procedures, signature binding needs supporting tools for
verification.
For example, PKI technology uses hashing algorithms and keys to demonstrate the
integrity of signed records. The digital signature is linked to the electronic record by
incorporating an instance of the record, into the signature itself. The link must be retained
for as long as the record is kept, which may be long after the signer has departed from the
company.
The records/signatures linking using password/ID code-based signatures is either
based on the use of software locks, on storing the electronic signature in a database table
Appendix C
196
separate from its associated record, or on storing the signature within the subject
electronic record. All these schemes require procedural and technological controls to
ensure that the requirements of Part 11.70 are met.
Appendix D
Criticality and Complexity Assessment1
PURPOSE
The Criticality and Complexity Assessment worksheet provides the guidance needed to
determine whether or not validation is required for new or modified computer systems. If
validation is required, the worksheet assigns it a level from 1 to 3. The validation level
determines the amount of documentation, testing, resources, and training that may be
required for the project.
There are five steps involved in completing the criticality and complexity assessment:
At the end of the assessment, the Deliverables Matrices detail the documentation that
may be required for each level of validation.
An outline of the criticality and complexity assessment results must be included in the
Requirements/Specification document for Validation Level 1, or the Validation Plan for
Levels 2 and 3. In those cases where the Requirements/Specification document and
Validation Plan are not required, an outline of the criticality and complexity assessment
result must be included in the qualification protocols.
The selection of deliverables for each validation level may differ from those indicated
in the Deliverables Matrices depending on the project documentation deliverables
policies used by each regulated company.
1
This Criticality and Complexity Assessment is taken from McNeil Consumer & Specialty
Pharmaceuticals (MCSP) computer systems validation methodology. Naming of deliverables in this
appendix is taken from MCSP methodology. Naming of deliverables in this book is not consistent
with MCSP methodology
Appendix D
198
VALIDATION MODEL DECISION TREE
Step 1: GxP Attribute Evaluation

Is the system being used to perform a regulated operation, or to feed information to
another system supporting regulated information activities? Does the computerized
system impact or affect any of the following attributes? If not, then it is not within the
scope of this assessment.
YES or NO
1.1 Does the system control a manufacturing and/or testing process
that has a direct impact on product quality? Product quality
attributes include identity, efficacy, strength, dosage,
safety, and purity.
1.2 Does the system produce data that is used to release
components or materials?
1.3 Does the system create, modify, maintain, archive, retrieve, or
transmit GxP-related information in electronic format that is
required by regulatory agencies?
If the answer to any of the questions in Step 1 was yes, then proceed to Step 2.
If the answers to all of the questions in Step 1 were no, then:
Appendix D
199
a. The computer system does not perform functions in a regulated operation and the
computer systems validation written procedures are not applicable.
b. If preferred, the computer system may be subject to an alternative SLC, one that is
not as rigorous as that required by the applicable regulation.
c. Attach support documentation with the evaluation, if applicable, and store in the
project file.
Proceed to Step 2.
Step 2: Criticality Assessment
What would be the impact on the regulated area if the data in the system were corrupted,
or if the system produced incorrect results? The higher the impact, the more extensive the
validation needs to be.
Criticality Matrix: What is the degree of criticality of the system? Enter either Low or
High in the Rating box as applicable.
YES
or NO
2.1 Has the system been implicated or found deficient by
any regulatory authorities?
2.2 Does the system (or should the system based upon
GxPs)
utilize any type of electronic signature?
2.3 Can the system be classified as any of the following:
Laboratory Data Systems
Adverse Event Reporting System
Supervisory Control and Data Acquisition System
Stability Reporting/Tracking System
Nonconformance/Investigation/Deviation System
Complaint System
Electronic Batch Record System
Calibration/Preventative Maintenance Tracking System
ERP/MRP
Label Management System
Facility Monitoring and/or Control System
2.4 Does the system or modification/change create, modify,
maintain, archive, retrieve or transmit data that has been
predetermined to be either a critical parameter or data
that
is used to release a batch?
2.5 Does the system or modification/change create, modify,
maintain, archive, retrieve or transmit data used in
regulatory submissions?
2.6 Does the system or modification/change dictate the
operation
of a critical process?
Criticality
Rating
Appendix D
200
High/Low
If the answer is YES to any of the questions 2.12.6,

then
the group is considered to be a high-criticality group.
If the answer is NO to all of the questions 2.12.6, then
the
group is considered to be a low-criticality group.
Proceed to Step 3.
Step 3: Complexity Assessment
How complex is the system? Would the user immediately detect any problem? For
example, if a system is producing three forms, and the user can immediately detect the
wrong form, the extent of validation may be minimal. Conversely, if the system were
performing complex calculations to determine if a product passes or fails product safety
tests, and there is no simple way to check these calculations, the extent of validation
would be greater.
Complexity Matrix: What is the degree of complexity of the system? Enter either Low
or High in the Rating box, as appropriate.
YES or
NO
3.1 Technology: Does the system or system modification/change
require the application of existing technology that is new to the
company or that involves the use or development of a new
technology that requires specialist skills and expertise?
3.2 Resources: Does the system or system modification/change to
be implemented require a dedicated team or a cross-functional team?
3.3 Software: Does the system or system modification/change affect
other subsystems? Is there a large volume of code, or are there
very complicated interfaces associated with the system?
3.4 Infrastructure: Does the system or system modification/change I
mpact multiple or company-wide functions, or is new
infrastructure required?
Complexity
Rating
High/Low
If the answer is YES to any of the questions 3.13.4, then the

system is considered to be a high-complexity system.
If the answer is NO to all of the questions 3.13.4, then the
system is considered to be a low-complexity system.
Proceed to Step 4.
Note: In the complexity matrix:
Technology refers to the level of technology (either process or equipment) required to
develop, implement, or change a system.
Appendix D
201
Resources refer to the people, equipment, and facilities/utilities required to develop,

implement, or change a system.
Software refers to the complexity of the Software Development effort required to
implement or change a system.
Infrastructure refers to the infrastructure (e.g., organizations, systems, and
procedures) required to implement, operate, or change a system.
Step 4: Determine the Required Level of Procedures
Enter the criticality and complexity rating determined in Step 2 and Step 3.
Criticality Complexity
Using the table below, determine the validation level.

Criticality Complexity Validation Level
Low
High
LEVEL 1
N/A
N/A
Low
High
LEVEL 2
LEVEL 3
Enter the Validation Level.

Level
Enter the Category of Software. Note: same as GAMP.

Category of Software
Proceed to the Deliverables Matrices that follow and select the matrix that is applicable
for the Validation Level that has been determined. This will enable the documentation
requirements to be determined. Then proceed to the page after the Deliverable Matrices
and enter the documents required for the computer system.
Deliverables Matrices
The following three matrices detail the deliverables for each level of validation. There is
one matrix per validation level.
Appendix D
202
Level 1: Deliverables Matrix
Phase
Documentation Requirements
Software Categories (GAMP)

1 2 3
4 5
Requirements Qualification and CSV Plan

Vendor Qualification Plan/ Certification
User Requirements
NR1* NR1 NR1
Functional Specifications
R1* R1 R1
Design
System Design Specifications
R2 R2
Traceability Analysis
Implementation Unit Test Documentation & Report
Code and Code Review
Documentation
In-process Vendor/System
Developer Audit Report
Integration Test Procedures
Test
Integration Test Report
IQ Protocol and Report
R3 R3 R3 R3 R3
OQ Protocol and Report
R3 R3 R3 R3
PQ Protocol and Report
Vendor Qualification Report
Validation Report
NR=Required for new systems
Rn=Required documents for group n may be combined into a single
document (i.e., one document for all R1 deliverables)
*=A comprehensive user manual may be substituted for the user
requirements and/or functional specification
Blank=Not required
LEVEL 2: DELIVERABLES MATRIX
Phase

User Requirements
Design
Code and Code Review Documentation
In-process Vendor/System

1 2 3
4
5
R1
R1
R1
NR1* NR1 R1
R1* R1 R1
R2 R2
R2
NR3 NR3
R3 R3
Appendix D
203
Test

R4 R4 R4
R4 R4
R4 R4
R4 R4
R4
R4 R4
Validation Report
R5
R5 R5
Rn=Required documents for group n may be combined into a single document (i.e., one
document for all R1 deliverables)
*=A comprehensive user manual may be substituted for the user requirements and/or
functional specification if approved in the qualification/validation plan.
Blank=Not required
Level 3: Deliverables Matrix
Phase

1 2 3
4 5
Requirements Qualification and/or CSV Plan

R1 R1 R1
Vendor Qualification Certification
R2 R2
(New/Modified Applications)
User Requirements
NR3* NR3 NR3
R3* R3 R3
Design
R4 R4
R5 R5
NR6 NR6
Code and Code Review Documentation
R6 R6
In-Process Vendor/System
R7 R7
Test
R8 R8 R8 R8 R8
R8 R8 R8 R8
R8 R8 R8
R9 R9 R9
Validation Report
R10 R10 R10
Rn=Required documents for group n may be combined into a single document (i.e.,
one document for all R1 deliverables)
*=A comprehensive user manual may be substituted for the user requirements and/or
functional specification if approved in the qualification/validation plan.
Blank=Not required
Appendix D
204
Deliverables Matrix:
Phase

12345

User Requirements
Design
Test Plan
Traceability Matrix
Code and Code Review
Documentation
In-Process Vendor/System
Test
IntegrationTest Report
Validation Report
Appendix E
Sample Development Activities Grouped by
Project Periods
Using a critical and complex custom-built system as an example, and the conventional
waterfall model (see Chapter 7, Figure 71) as the basic framework for specification,
design, and testing development methodology, the following sections contain a detailed
description of a development methodology grouped by periods. An overview of this
appendix is provided in Figure E3.
CONCEPTUALIZATION PERIOD
The system life cycle starts with the conceptualization period. This period is critical and
is the highest level of design, and is primarily the responsibility of the system owner. The
purpose of the conceptualization period is to establish an integrated model of the
operation to be automated or re-engineered. This model provides the criterion to
company senior management, enabling them to make a business and/or regulatory
decision on whether to continue with a project.
The users needs are analyzed in order to determine appropriate prospective solutions.
The prospective solutions are high-level solutions identifying major constraints, system
interfaces, user interfaces, and proposed technologies. These solutions are also used to
assess the criticality and complexity of each proposed system component (refer to
Appendix D for more information). This assessment will provide information on the
resources and documentation requirements for the project.
During the conceptualization period, the scope of the system is defined based on the
strategic objectives contained in a marketing proposal or similar document.
Supporting information is obtained for this period from interviews and by reviewing
documentation generated by operations, R&D, QA, and Engineering/IT. Part 11 should
be included as part of the integrated model of the operation to be automated or reengineered.
During the operational life of the computer system, a change request to enhance or
amend the computer system will result in this period being initiated again and the original
concept being reviewed and updated if necessary.
The outputs of the conceptualization period can include operational checks/sequencing
models. A new or updated concept document is produced during this phase.
Appendix E
207
DEVELOPMENT PERIOD
The purpose of the Development Period is to define a structured process to specify,
design, build, test, install, and support the business operation to be automated or updated.
The output of the development period may include the following:
plans (e.g., project, validation)
written and approved requirements (e.g., users, functional, design) that describe what
the application is intended to do and how it is intended that it should do it
test results and an evaluation of how these results demonstrate that the predetermined
requirements (e.g., users, functional, design) have been met
system acceptance for use
Requirements Gathering and Specification Development
The scope of the system generates many of the support requirements for the operations,
usually provided by the system owner. The term requirement defines a bounded
characterization of the scope of the system, and contains information essential to support
the operation/operators. Some of these requirements include functional capacity,
execution capability, operational usability, information needed to support validation,
installation, commissioning, SLC documentation required, users manuals, training,
maintenance manual, system maintenance, system test plan, acceptance criteria, and
regulatory compliance (e.g., Part 11).
The majority of computer system developments fail due to poor gathering of
requirements. This situation negatively affects subsequent development activities and
associated working products.
The system requirements are captured in the requirements specification deliverable,
which describes what the system is supposed to do from (a) the process, (b) the users,
and (c) the compliance perspectives. The requirements specification deliverable may be
used as a framework to select the supplier/ integrator and to develop the PQ protocol.
The requirements specification deliverable must include an overview of the process in
order to familiarize the application software developer with the user, process and data
acquisition requirements of the system, as well as special considerations for the project.
The system functionality must be well defined at the outset, in order to provide the
prospective supplier/ integrator with enough information to provide a detailed and
meaningful quotation.
Specifically, on data acquisition systems, the requirements specification deliverable
must include definitions of (i) the data to be collected, (ii) how the data will be used, (iii)
how it will be stored and retention requirements, (iv) data security requirements, and (v)
where each operation will be completed. The requirements specification deliverable
addresses the following:
the scope of the system and strategic objectives
the problem to be solved
Appendix E
208
process overview, sequencing requirements, operational checks

sufficient information to enable the supplier/integrator to work on a solution to the
problem (e.g., device-driven sequencing, the methods required of the presentation of
data, data security, data backup, data and status reporting and trending, etc.)
redundancy and error-detection protocol
operating environmental
interfaces (e.g., to field devices, data acquisition systems, reports, and HMI), I/O lists,
communications protocols, and data link requirements
information gained from operators and supervisors on the system design requirements
and expectations in order to influence how the system is designed and operated
type of control and process operations to be performed
data storage requirements
transaction/data timing requirements and considerations
regulatory requirements (e.g., validation, Part 11)
preliminary evaluation of the technology
feasibility study and preliminary risk assessment
safety and security considerations
Part 11, security, other requirements
nonfunctional requirements (e.g., SLC development standards, programming language
standards, program naming convention standards, etc.)
Each requirement in the requirements specification deliverable must have the following
attributes:
Unambiguous All requirements must have only one interpretation
Verifiable
A human being or a machine must be able to verify that the system
correctly implements the stated requirements
Traceable
The requirements must be traceable from other design and test
documents
Modifiable Unanticipated changes must be able to be made easily
Usable
The URS must be usable by not only the development team but also
subsequent maintenance teams who will be called upon to modify
and change the system
Consistent
Individual requirements must not conflict with each other
Complete
It must contain clear descriptions of all features and functions of the
system. It must clearly contain definitions of all known situations
the system could encounter
The requirements specification deliverable is tested in an operating environment during

the PQ and must include the verification of the procedural controls associated with the
system, as identified in the requirements specification deliverable.
The following section provides an example of the general content of a requirements
specification deliverable and the type of information to be included.
Functions
This subsection defines the required system functions, modes of operation, and
behaviour. It should address the following:
Appendix E
209
functions required. Information on the process or existing manual system should be

included here, if not covered adequately elsewhere
calculations, including all critical algorithms
modes of operation (e.g., start-up, shutdown, test, fallback manual, automatic)
performance and timing requirements. These should be quantitative and
unambiguous
failure response. The action required in case of selected software or hardware failure
safety and security
Data
This subsection states the data-handling requirements. It should address the following:
definition of the data, including identification of critical parameters, valid data
ranges and limits, and how out-of-range data should be handled
capacity requirements
format requirements (e.g., when transferring data between systems and for meeting
Part 11 data inspection requirements)
access speed requirements
archive requirements
Interfaces
This subsection defines any system interfaces and contains the following subsections:
interfaces with users. These should be defined in terms of roles (e.g., plant operator,
warehouse administrator, manager, etc.) and/or functions as appropriate
interfaces with other systems
interfaces with equipment (e.g., sensors/actuators). This includes I/O listings for
process control systems
Environment
This subsection defines the environment in which the system is to work. It contains the
following subsections:
layout. The physical layout of the plant or work place may have an impact on the
system (e.g., long-distance links, space limitations, proximity to electrical noise
sources, etc.)
environmental conditions (e.g., temperature, humidity, dusty or sterile environment,
etc.)
Constraints
This section defines the constraints on the specification of the system. It contains the
following subsections:
timescales and milestones, as appropriate
compatibility. This takes into account any existing systems or hardware, and any
company or plant strategies
Appendix E
210
availability. This states the reliability requirements, and the maximum allowable
periods for maintenance or other downtime
procedural constraints (e.g., statutory obligations, legal issues, working methods and
user skill levels)
maintenance (e.g., ease of maintenance, expansion capability, likely enhancements,
expected lifetime, long-term support)
Life cycle
This section defines any requirements concerning the development life cycle. It contains
the following subsections:
development (e.g., procedures for project management and quality assurance,
mandatory design methods, design review requirements, etc.)
testing (e.g., special testing requirements, test data, load testing, simulation,
reporting of the test results, etc.)
Deliverables
This section defines what deliverables are required for the project. It should address the
following:
how deliverable items are to be identified
in what form deliverables are to be supplied (e.g., format and media)
documents. What the supplier is expected to deliver (e.g., functional specification,
testing specifications, design specifications, design reviews, test results, etc.)
data to be prepared or converted
tools
training courses
archiving facilities
Support
This section defines what support is required after acceptance
Glossary
This section contains definitions of terms, which may be unfamiliar to the reader of the
requirements specification deliverable
Appendices
Any additional information that may help the computer system supplier/developer to
provide a quotation can be provided as appendices
Develop a Validation Plan
The validation plan is discussed in Chapter 8.
Appendix E
211
Conduct a System (Hardware and/or Software) Risk Analysis

Software development is one of the most risk-prone management challenges. Risk factors
can negatively impact the development process and, if ignored, can lead to project failure.
To counteract these factors, system risk must be actively assessed, controlled, and
reduced on a routine basis.1
Risk is defined as 'the probability of an undesirable event occurring and the impact of
that event if it does occur.' The result of this analysis will influence the degree to which
the system development, implementation, and maintenance activities are performed and
documented. By evaluating the system risk analysis, the system owner may uncover
potential problems, which can be avoided during the development process. The chances
of a successful, if not perfect, system implementation are improved.
There are seven steps necessary for assessing and managing risks:
a description of the risk
the cause(s) of the risk
the level of concern regarding the risk based on a qualitative estimate, including any
definition of terms used
the likelihood of occurrence of the risk based on a qualitative estimate, including the
definition of terms used
the method(s) of control used to eliminate or mitigate the risk, e.g., a change in the
design specification, alarm warning and/or error messages, implementation of a
manual process and/or workaround
the traceability of each method of control to the corresponding critical requirement
specified in the requirements specification deliverable
monitor and manage the implementation of the requirements identified during the risk
analysis
Some of the items to consider when conducting the risk analysis include:
regulatory areas of the system
size of the computer system and number of users
complexity of the system in terms of the design
type of data the system handles
functionality of the system
interaction with and impact on existing systems
vendor reliability
effect/impact of system noncompliance
safety to the customer, employee, and community
cost versus benefit of the system
1
Guidelines for Successful Acquisition and Management of Software Intensive Systems, Vol. 1,
Software Technology Support Center, DAF, February 1995.
Appendix E
212
The results of the risk analysis review must be documented. To adequately analyze the
risks, a comprehensive requirements specification deliverable is required. The best time
to perform an initial risk analysis is immediately after the project team has completed the
review of the requirements specification deliverable. The review of the solutions agreed
upon and implemented as a result of the risk analysis should be performed during a
technical design review.
During the design and development of computer systems or application software to be
used as components in medical devices, to be manufactured and/or sold in the USA,
consideration must be given to meeting the requirements for design in 21 CFR Part 820
of the FDA regulations.
Conduct a Supplier Audit of the Computer Technology Suppliers
and/or Developer
Develop a Computer System Specification
The computer system specification includes information on how the operation can be
controlled using the features in the selected application package. A review of the process
must be conducted to familiarize the computer technology supplier and/or developer
(external or in-house) with the requirements defined in the requirements specification
deliverable. This system view is essential, particularly when the software must interface
with other elements such as hardware, people, and databases. The review of the process
includes the further gathering of requirements at the system level, and top-level design.
The computer system specification deliverable for the system identifies the detailed
requirements to be incorporated into the design and tested during the FAT, SAT, and/or
OQ. The computer system specification takes the information provided by the
requirements specification deliverable, and provides the following:
functions defined in detail
interfaces further described
design constraints and performance characteristics established
database characteristics identified
test and validation criteria defined
The documentation generated during this phase may include data flow diagrams, data
definitions, detailed descriptions of the processes and functions, performance
requirements, and security considerations.
The elements of the computer system specification deliverable relate directly to the
documentation and test cases contained in the OQ protocol. Once the computer system
specification deliverable is approved, it is possible to start the detailed design phase and
to generate the FAT, the SAT, and/or the OQ documentation.
The following is an example of the general content of the computer system
specification deliverable for a system. If no requirement has been specified in the
Appendix E
213
requirements specification deliverable for a particular section, the section must still be
included but must state Not Applicable.
Introduction section
This section contains the following information:
who developed the computer system specification deliverable, under what authority,
and for what purpose
relationship to other documents
Overview section
This section states the essential system functions, and interfaces to the outside world. It
should cover the following:
key objectives and benefits
high-level description. This should provide breakdown of the computer system into
the main subsystems (hardware and software)
the main interfaces between the system and the outside world
assumptions. This should state any design or implementation assumptions (e.g.,
standard packages, operating system, hardware)
nonconformance with the requirements specification deliverable. Any divergence
between the computer system specification deliverable and the requirements
specification deliverable must be clearly noted
Functions section
This section takes the high-level description, and breaks it down to the level of the
individual functions. It describes the functions and facilities to be provided, including
specific modes of operation. This description is based on the features implemented in the
selected application. The following aspects should be addressed:
the objective of the function or facility, and the details of its use, including the
interface with other parts of the system. Critical calculations and algorithms should
be highlighted
the performance (e.g., response, accuracy, and throughput), which should be
quantitative and unambiguous
safety and security. The topics covered may include: the action to be taken in case of
selected software or hardware failures, self-checking, input validation, redundancy,
access restrictions, time-outs and data recovery
functions that are configurable and any limits within which the configuration can
take place
clause level traceability to specific requirements in the requirements specification
deliverable
Data section
This section defines the data on which the system is expected to work. The
following aspects should be addressed:
Appendix E
214
definition. The data should be defined in a hierarchical manner with complex objects
being built up from simpler objects (e.g., files of records; complex data types
defined in terms of simple data types). Critical parameters should be highlighted
access (e.g., which subsystems need read or write access to each data item, access
method, speed and update time, read/write interlocks)
data capacity, retention times, and details of how the data should be archived
Interfaces section
This section defines any system interfaces. It contains the following subsections:
interface with users defined in terms of roles (e.g., operator, administrator, clerk, and
system manager). Topics to consider include: the facilities available, suitable types
of peripherals, the general format of displays and reports
interface with other systems. Topics to consider include: data transmitted and
received, rate of transfer, data protocol, and the security of the interface
interface with equipment (e.g., sensors, plant equipment). Topics to consider
include: data transmitted/received, data format, validation, and error checking
Nonfunctional attributes section
This section defines how the system will meet nonfunctional requirements. It
contains the following subsections:
availability (e.g., reliability, redundancy, error checking, standby operation)
maintainability (e.g., expansion and enhancement possibilities, spare capacity, likely
changes in environment, system life expectancy)
Glossary section
This section contains definitions of terms, some of which may be unfamiliar to
the readership of the computer system specification deliverable.
Appendices
Where appropriate (e.g., small systems), appendices may be provided to define
hardware and software specifications.
Develop a Technical Design Specification
The design process and resulting deliverable are based on, and clause traceable to, the
information in the requirements specification deliverable. The technical design
specification must specify how the combined hardware and software products will satisfy
the requirements.
Also specified in the technical design specification deliverable are the subsystems,
components and interfaces, data structure, design constraints, algorithms and system
decomposition of the computer system to be supplied by the project. The objective of a
good modular design or partitioned design is to separate both data and functions into
modules. Each module is called a unit. This concept is very important when linking the
designed units and when testing these units. Modular design is also a key concept when
developing computer systems using an objectoriented development methodology.
Appendix E
215
The technical design specification deliverable for the computer system is a document
defining the system in sufficient detail to enable it to be built, and to enable code to be
produced for it, including:
the system architecture and functions
hardware/software installation specifications and configuration
components
modules
interfaces (machine and human)
drawings or pictorial representation of the hardware that associates part of the system
and its interfaces to other hardware
system operation under both normal and abnormal conditions
logical and physical structure of the software programs
safety and security
design and programming standards
The technical design specification deliverable may include block diagrams, flow
diagrams, state transition diagrams, loop control calculations, and narrative descriptions.
Depending on the technology available and the cost, automated functions can be
assigned to the computer hardware or software. The computer hardware can be further
decomposed into a number of subelements.
The software design concepts such as modular, top-down, and decomposition of
the design requirements have been used in the computer science fields for over 20 years.
However, the application of these concepts can vary.
The support and maintenance of the software is simplified when consistent formats
and techniques are utilized. Troubleshooting programming logic not implemented in a
structured format is frustrating and tedious.
Developing a system that provides for ease of operation and maintenance requires that
the software engineers, user owners, and end-users agree on the:
design of the operator interface including overview displays
format of the program documentation and operational manuals
partitioning of each program into simple modules that can be reused throughout the
overall application software
It is imperative that an agreement is reached on a structured approach before any
programming/configuration of the software begins. This approach should provide the
basis for a design specification deliverable which defines the system development scope.
Typically, 80% of the programming or configuration process can be partitioned into
modules that are reusable throughout the system. Once a module has been developed and
tested using simulation techniques, it can be reused with little risk. It is important that
these modules are well documented, so that persons wishing to make use of the module
can easily understand them. It is important that the misapplication of modules is
minimized.
The remaining 20% of the programming/configuration process may represent custom
logic unique for that particular area of the application. Custom logic can be made easy to
understand and maintained through the use of consistent formats, variable utilization, and
effective documentation within the logic.
Appendix E
216
When a structured and modular architecture is used, the software becomes easier to
read, troubleshoot, modify, and extend. It quickly becomes obvious that a structured
approach is vital in these flexible yet complex systems.
The logical structures or implementation contained in the technical design
specification deliverable must be reviewed during a design review. A design review is
performed based on an approved procedure. The objective of the design review is to
verify that the logic structure including the modularity, reuse, functionality, and
maintainability of the implementation, and the clause traceability of the computer system
specification deliverable to the technical design specification deliverable. The technical
design specification deliverable also contains information on how the risks inherent in the
system design were resolved.
The system design is released for implementation once the technical design
specification deliverable has been approved. All changes made to the computerized
system design after the approval of the technical design specification deliverable are
reviewed and documented using a change management system/control.
During each software or system development step, the developer provides documented
evidence that implementing the requirements specified in the requirements specification
deliverable developed the product. During the design, the in-process (internal) audit must
be carried out in order to verify that the design of the computer system satisfies the
requirements described in the computer system specification deliverable, and that the
code has been developed in accordance with the technical design specification
deliverable.
A comprehensive design review or design qualification should be performed to
evaluate the design capabilities to fulfill the requirements, and a business case design
solution prepared.
The physical implementation of the system described in the technical design
specification deliverable must be verified during the IQ.
Design key practices can be found in Appendix I.
Consideration of the Human Factors in the Design of Operator
Interfaces
Poor operator interface design induces errors and inefficiency among even the besttrained operators, especially under conditions of stress, time constraints, and/or fatigue.
Although labeling (e.g., user documentation) is extremely important for good
performance, even well-written instructions are cumbersome to use in conjunction with
actual operation. In addition, it is difficult to write coherent documentation that describes
awkward operating procedures.
Although both hardware and software design influence the operators performance, the
logical and informational characteristics provided via software are increasingly crucial.
Data presented in an ambiguous, difficult-to-read, or nonintuitive manner poses the threat
of an incorrect reading, misinterpretation, and/or improper data entry. An example of this
could be a crowded display with cryptic identifiers, combined with a time lag between
the operator response and the displayed responses. Such design characteristics overtax the
users abilities (e.g., memory, visual perception, decision-making, etc.), and resultant
errors may have serious consequences.
Appendix E
217
Software Development
This is the lowest level of requirements decomposition. The technical design
specification deliverable is translated into software source code. Source code is the
human readable version of the list of instructions (programs) that enable a computer to
perform a task. Source code is written in the programming language appropriate for the
application to be generated.
Except for a very small number of programming languages, source code cannot be
understood and executed by a computer system. Consequently, the source code is
converted into a machine-executable form prior to being run. Machine-executable code is
also known as an object program. This conversion into machine code is called
translation, performed by another program called an assembler, compiler, or interpreter,
depending on the programming language involved.
All these activities are directed by procedures containing programming standards,
naming conventions, and the configuration settings for the compilers, assembler, and/or
interpreters. In addition to the object program, the compilers/assembler used will list the
results of the translation process, highlight any errors that may have been encountered,
with possible resolution. An interpreter does not create any files or lists.
In a mature software development environment, software is developed as individual
components or small units. Each unit is tested individually and integrated with other units
to form a module. The accurate implementation of each module is achieved by following
the technical design specification deliverable. The output of the software development
process is an assembled system.
The software development process involves coding, debugging, and formal testing
(code audits, unit testing, and integration testing) by the developer. The developer must
comply with the established programming standards, where applicable, to ensure that the
program has a uniform content and to assist with the ongoing understanding of the source
code.
Code reviews, inspections, and audits
During and after the coding process, code reviews, inspections, and audits on
application software and/or configurable products must be performed. Refer to
Chapter 9. The objectives of code reviews, inspections, and audits are:
the verification of the absence of dead code (CPG 7132a.l5)
confirmation of a modular design
verification of the conformance with the programming approach selected
verification of the design modification that were implemented as a result of the risk
analysis
the features incorporated into the code in order for it to be efficiently maintained
during its operational life
Code reviews, inspections, and audits may make use of high-level system data flow
diagrams (e.g., transaction flow diagrams) in order to determine the scope and depth of
the reviews, inspections, and audits.
Appendix E
218
Conduct software module (unit) testing and integration testing

Figure E1 depicts the essential activities of unit, integration, and system level (FAT,
SAT, OQ, PQ) testing, and the relationships between them. These activities are discussed
in the following sections.
A unit test is a white box-oriented test, conducted on modules that have been
structurally tested. This type of test involves the standalone software unit before it is
integrated into a complete system or subsystem. The test cases are designed to look for
errors in the logic design, data processing, and the statement of execution. Uncovered
errors are removed, and the test case that located the error, and any other associated test
case, re-executed.
It is not always feasible to test a program with real equipment or hardware. In these
cases, a program can be put under the required testing conditions and tested using input
simulation.
Figure E1. The relationship

between the testing activities.
The items that can be checked during unit test are as follows:
that it functions as specified in the appropriate design document
maximum or minimum limits testing
reaction to incorrect inputs
exception handling
algorithm assessment
adherence to documented standards
Appendix E
219
The software integration testing is an orderly progression from unit testing and
incremental testing in which software, hardware elements, or both are combined and
tested, until the entire computer system has been assembled.
The objective of the integration test is to uncover errors associated with software
interfaces and/or the integration of the software with the hardware. Unit-tested modules
are built into a program structure that has been dictated by the design documentation.
The characteristics to be evaluated during the software integration testing are:
interface integrity
functional validity
information content
The software test strategy and detailed test scripts (protocol) are reviewed and approved
by the Project Team. This review will verify that the test cases will stress the program
interfaces, challenge the data boundaries and limits, and verify that the test cases are
traceable to the appropriate clauses in the requirements specification deliverable. The
review also verifies that the test cases adequately cover the risks identified in the Risk
Analysis. The unit/ integration test results are documented by the programmer, and
reviewed, approved, signed, and dated by QA.
Preparation of the Qualification Protocols
Qualifications are formal examinations of the completed system in order to determine and
record the acceptability of the design, fabrication/development, and integration of the
hardware/ software and associated utilities.
Qualification protocols are written documents prepared before conducting the
qualification, describing the features of a particular application or item and how it should
be tested. Qualification protocols identify the objectives, methods, and acceptance criteria
for each test function contained in the applicable specification deliverable; and identifies
who is responsible for conducting the tests. In addition, the protocols should also specify
how the data is to be collected, reported, and analyzed to determine if the acceptance
criteria were met. The protocol should be reviewed by personnel with an appropriate
understanding of computer systems and the functionality of the indicated system.
Following the review, qualified personnel must approve each protocol in accordance with
the company quality assurance procedures. In the software engineering world,
qualification protocols are equivalent to test procedures. Traditionally the scope of the
qualification protocols encompasses the following:
installation
operation
performance
The focus of qualification of focus protocols is always prospective and the detail
contained in each dependent upon the validation tasks performed.
The contents of the FAT, SAT, and OQ tests are similar. The differences between
these activities are the test environment and the formality of the testing. Figure E2
depicts an example of a qualification timeline that contains these activities. FAT, SAT,
and OQ are described in Chapter 10. The successful execution of the protocols provides
Appendix E
220
documented evidence that the computer system is performing as specified in the

requirements specification deliverable, computer system specification deliverable, and
the technical design specification deliverable.
Figure E-2. Qualification Timeline.

The recommended contents of a protocol are listed below. The presence or exclusion of a
particular item depends on the scope of the computer system validation.
a. Purpose or objective of the protocol
b. Scope of the validation activities
c. Definition of the system to be validated
d. Responsibilities for:
Creation of the protocol
Review/approval of the protocol
Execution of protocol
Review of the validation effort required
Approval/acceptance of the validation effort
The meaning of each signature
e. Validation rationale (including references to any company policies):
Discussion of the regulatory nature of the computer system
Reference to the company validation policies, methods, or standards
Strategy for the validation (e.g., phased)
f. Exclusions and assumptions
g. High-level description of the system:
Hardware components (the computer platform and any equipment that interfaces to
or is controlled by the system)
Software components:
System software (operating system and utilities)
Noncustom-supplied software
Custom software
Appendix E
221
Network/communications software.
Interfaces to other systems or software
High-level diagram of the system hardware configuration (if applicable)
High-level diagram of the system functions (if applicable)
h. Prequalification documentation verification:
Statement of work or project initiation
Business, system, and functional requirements specification deliverable
Detailed system design specification deliverables (computer system and technical
design specifications)
Development testing specification deliverables and associated results:
Structure/code evaluation (e.g., code review records)
Unit testing
Integration testing
System testing
Reference to any project or development standards and methods
i. Supplier qualification (if applicable)
j. System physical location/environment and requirements.
k. Hardware IQ:
Operating environment and parameters
Hardware components
Interfaced equipment (if applicable)
Network/communications hardware (if applicable)
l. Software IQ:
Operating system and utilities software
Noncustom-supplied software
Custom software
Network/communications software (if applicable)
m. OQ
n. PQ
Test plan (if size of project warrants a separate document of this type)
Verification of traceability to user requirements
Criteria and procedure for the identification and resolution of nonconformances that
occur during testing
o. Security (physical, platform, and application)
p. Training
q. Operational procedures:
System usage
System support, maintenance, and monitoring
Appendix E
222
r. Development of a Requirements Traceability Matrix

s. Change and configuration management
t. Backup and restoration
u. Disaster recovery and alternative procedures
v. Incident reporting for the computer system (hardware and software) once it has been
used in a production environment
w. Periodic review
x. Acceptance criteria for the successful completion of the system validation
y. Addenda, supplements, or revisions to the protocol
z. Validation report
a. Review and approval of the protocol for execution
Since the protocol is prospective in nature, sections h through v contain the following
information in each section:
Identification of the function being tested or a unique test identification number or
indicator
Any prerequisites necessary to perform the test
If using an electronic test tool, indicate the test script, electronic test file name, or other
identifier
The test script, which includes:
Steps or actions to be performed
The input data for the appropriate steps
The expected result or expected system response to the input
An area for entry of the actual test result or for entry of directive information that
references an attached output (for example, an attached report, screen print, audit
log, output from an automated test tool, etc.)
Areas to indicate pass or fail for critical steps and initials of the tester
The acceptance criteria for the entire test script
Area for the date of the test
Area for signature of tester
Area for indication of pass or fail of the test script
Area for nonconformance comments
Area for the date reviewed
Area for the signature of the reviewer
Area for comments regarding any nonconformance resolutions
Hardware Integration and Program Build
Hardware integration involves the connection of hardware and network components
based on the system hardware configuration. This configuration is specified in the
computer system specification deliverable and further defined in the technical design
specification deliverable. If the hardware integration is performed at the
suppliers/integrators site before the FAT, there is no need for a formal document to
record the hardware installation. If the hardware integration is performed at the target or
Appendix E
223
production environment, then formal documentation is required. In the context of FDA

regulated industries, the formal document is called a hardware IQ.
Conduct a Factory Acceptance Test
The FAT, Figure E2, is a mutually agreed acceptance test of the system between the
supplier/developer and the user. It is performed by the computer technology
supplier/developer at the suppliers/developers site. The user will often witness this test
to accept the system for delivery and to mark a contractual milestone.
The SAT has traditionally been seen as engineering documents and activities in
support of the commercial and technical contracts, used as risk-reduction steps as
opposed to qualification activities. As computer technologies suppliers and contract
developers become better at working in a controlled environment, the complexity and
coverage of the acceptance testing is improving. This is reflected in the current GAMP
guide (GAMP4 2001), where qualification activities are shown as being completed
during acceptance testing activities. As referenced in Sidebar 101, the FAT has the
possibly of being OQ1, often a positive way forward as it reduces the amount of rework
associated with testing/qualification activities. Obviously there are going to be some
drawbacks using this approach, and is a balance to be evaluated thoroughly before
starting any project or project phase.
The plus points include:
qualification starts early and some elements can be completed in the
suppliers/developers facility. This could save on rework costs
reduced effort of retesting
modular testing with the appropriate level of rigor is possible
For example a system can be taken through all its exception handling routines in a
simulated environment; this might be less practical on a live system.
The minus points include:
acceptance testing has to fall under the appropriate level of change control to reflect that
it is a qualification activity
testing and qualification planning is more complex, as the full system coverage has to
be demonstrated
content and documentation standards of the acceptance testing documents need to be
capable of supporting qualification and need to be approved by QA
The chosen route should be planned accordingly and included as part of the Validation
Plan.
Typically, the FAT addresses the following:
interfaces
system functionality (e.g., sequences, limits, alarms, boundaries)
system operability and performance
initial user training, if applicable under the contract
critical parameters
Part 11 technical controls
Appendix E
224
electronic records management

logical security
management and security of user identification code and password (authentication
and/or electronic signature)
generation of accurate and complete copies of electronic records
system date and time integrity and security
audit trails, metadata, display and reporting of audit trails
operational, authority and device checks
signatures management (nonbiometrics/biometrics)
signatures manifestation
multisigning
signatures binding
signatures sequencing
The FAT provides evidence that the hardware and software are fully integrated, that they
operate as indicated in the computer system specification deliverable, and meet the
expectations of the user as defined in the requirements specification deliverable. This
final formal integration test should be completed in an environment very similar to the
operational environment. The system can be subjected to a real-world environment by
using emulators and/or simulators which mimic system interfaces. The users
representative should evaluate the supporting documents, the operation, system
functionality, and system reliability.
If the FAT results and audits conform to the contractual agreements the system is
accepted for shipment and installation. In particular, all identified critical problems and
deviations found during the execution of the FAT must be corrected before shipment.
CONDUCT A SITE ACCEPTANCE TEST

The SAT is an acceptance test comprising system installation, start-up, operational
testing, and handover to customer, performed by the supplier/ integrator in the
operational environment. As with the FAT, the users representative should witness the
tests and evaluate the results of the SAT.
The testing carried out in the development environment is not sufficient to verify the
integrity and performance of the system. The application software must also be tested
under actual conditions of intended use in the target or production environment. If the
application is to be supplied by a third party in accordance with the end-users
requirements specification deliverable, it should be possible to carry out a completely
prospective validation, as would be the case for a system developed in house, providing
the requirement for validation is identified at the beginning. As with all systems, the
responsibility for validation lies with the systems owner.
Early agreement should be reached between customer/owner and supplier with regard
to the documentation necessary to establish the validity of system parts under supplier
control. In addition, the supplier should agree to inspection(s) by the owner, or if this
would compromise the suppliers commercial interests, by an independent expert,
employed by the owner.
Appendix E
225
During the SAT, many of the FAT verifications are re-executed. The only difference
between the FAT and the SAT is that the FAT is performed under nonoperational
conditions, whereas the SAT is performed under operational conditions. Depending on
the configuration management controls established during the FAT, the field independent
FAT verifications and associated test case results may not be repeated during the SAT.
One key test in the SAT is the configuration and qualification of the hardware.
Depending on the type of computer system involved, this is possibly the first time that
the complete system hardware has been assembled. The hardware configuration provides
an equipment installation, configuration, and inventory list for each device and/or item of
equipment associated with the computer system and network hardware. The suppliers
manuals for each piece of hardware must be provided. The hardware qualification
includes:
system description and schematic drawings
instruments list/schedules and drawings
computer system hardware list (including part/serial numbers), purchase orders, and
requirements/data sheets
I/O lists/schedule
wiring checks
loop diagrams, cable schedules and routing diagrams, panel and junction
box/termination layouts
full calibrations certificates that list the test equipment and its traceability to national
standards, and include the limits of uncertainty results
verification of hardware integration
security verification
execution of system diagnostics
control equipment configuration verification
security verification (physical)
control equipment interface verification
During the software IQ the developer will produce a plan for installing the software in the
target environment as specified by the technical requirements specification deliverable.
The resources and information necessary to install the software product must be
determined and be made available. The developer must assist the system owner and
maintenance personnel with the setup activities: where the installed software product is
replacing an existing system, the developer must support any parallel operation activities
required by the contract. The installation plan and procedures must be documented.
The developer must install the software product in accordance with the installation
plan and procedures. It must be ensured that the software code and databases initialize,
execute, and terminate as specified in the technical design specification deliverable. The
installation events and results must be documented.
The software IQ includes:
man-machine-interface installation verification
database and configuration verification
version verification for all software used by the computer system
a list of the individual manuals that are applicable to the computerized system
Appendix E
226
a list of all computer software for specific processor modules associated with the unit;
this will include the source code, as applicable, and samples of the graphic displays for
various situations
workstation system configurations
a text copy of the latest revision of each application program
a copy of the module configuration database form for each module configuration type
a computer printout of the final field device point configurations
The software OQ includes all of the factory acceptance testing and, specifically, the
following:
A fallback plan must be devised for recovery of existing services in the event that
introduction of a new system or new functionality to an existing system causes service
degradation or interruption
If database conversion, migration, or preloading with data has to occur prior to OQ
testing, the verification of these activities and their associated data may be addressed
in either a pre-OQ test or during the OQ. Any data migration tools should be written
under the control of a software quality management system. The data migration
process should be verified and the verification performed before final data loading to
the new system
The Part 11 technical controls tested at the FAT
All major functions and interfaces must be exercised
Catastrophic recovery verification
Worst-case conditions for the application software (Note: As defined by the FDA the
worst case is a set of conditions encompassing upper and lower processing limits and
circumstances, including those within SOPs, which pose the greatest chance of
process or product failure when compared to ideal conditions)
Manmachine-interface functional verification
Boundary values
Invalid values
Special values
Decision point and branch conditions
Tests on the generation of reports
Hardware redundancy verification
Data acquisition and database access controls
Trending, alarm and event processing
Interface with other applications
Timing (if applicable)
Verification of security protection mechanisms (Logical)
In situations where the SAT is conducted according to approved protocols, the test results
are documented properly and reviewed, approved by the user and QA, it may not be
necessary to repeat the SAT tests during the IQ and OQ. The IQ and OQ should still
identify all of the necessary tests, but result areas can simply reference the appropriate
test carried out during the SAT, providing evidence of successful completion.
Appendix E
227
The IQ must identify the hardware, software, and interfaces used by the computer
system and demonstrate their installation, in accordance with the design drawings,
specifications, and supplier recommendations. These requirements are established and
documented in the technical design specification deliverable.
The OQ must demonstrate that the functionality of the system is as established in the
computer system specification deliverable. During the OQ, all systems interfacing with
the computer system must be operational throughout the entire process. Document
evidence must be obtained verifying execution of the various steps and functions detailed
in test. The protocol must be approved and completed by the appropriate personnel.
Conduct a Performance Qualification
The computer system needs a comprehensive PQ test designed to determine the degree of
accuracy with which the system conforms to the requirements specification deliverable.
The computer system PQ should not be confused with the process PQ and the product
qualification as described in the FDA Guideline on Principles of Process Validation. The
PQ occurs under operational conditions but not as part of the actual operational process.
The specific test cases include:
Verification of the system documentation, all procedural controls, migration, and
production implementations. Procedural controls include Part 11 controls not covered
by the technical controls
Determination of the accuracy of the system in receiving, recording, storing, and
processing the electronic records
Determination of the system accuracy in arriving at the appropriate disposition decision
based upon the data received by the system
Determination of the integration of the computer system into the operating environment
(refer to Chapter 2, Figure 21), including technologies, procedural controls,
controlled processes/operations, people, and documentation
Operational checks relating to electronic signatures, if applicable
The PQ must cover the full operating cycle of the computer system and operating process
and be performed following an approved protocol.
Prepare Qualification(s) and Project Report(s)
Once the protocols have been completed, the test results and data need to be formally
evaluated. The written evaluation needs to be presented clearly, in a manner that can be
readily understood. The report should also address any nonconformances or deviations to
the validation plan encountered during the qualification and their resolutions. The format
of the report should be similar to the structure of the associated protocols. The
qualification testing should be linked with the acceptance criteria in the relevant
specification deliverable, so that the PQ will link with the requirements specification
deliverable; the OQ will link with the company system specification deliverable; and the
IQ will link with the technical design specification deliverable.
Appendix E
228
In addition, all the document inspections and technical review results for all the
technologies used for the systems must be evaluated. A summary of nonconformances
and deviations, their resolution and impact, must also be included.
For very large validation projects, the project report should reference (by title and
document reference number) other documents satisfying protocol requirements. In
smaller validation projects, the actual evidence can be incorporated into the project as
appendices.
Documents associated with the system qualification and their results must be
assembled and reviewed by appropriately qualified personnel following the review, by
QA, and by the personnel appropriate for the level of criticality of the system.
The approval of the qualification reports provides confirmation that the computer
system as a whole is fit for its purpose, and that all essential documentation is available.
For computer systems controlling manufacturing equipment (e.g., process control
systems), the approval of the qualification reports indicate that the computerized systems
is released for Process/Product Performance Qualification. On other computer systems,
the approval of the reports releases the system to the user.
The test results for compliance with Part 11 must be included in, or referenced by the
associated qualification report.
The project report summarizes the outcome of each activity performed to develop or
maintain the computer system and the verification that critical checkpoints were reached
during the entire development process. It also verifies that the quality development
procedures identified in the project plan were complied with.
All verification and testing results obtained during the project must also be addressed
in the project report. The approval of the project report enables the computer system to be
released for use.
Conduct a Process/Product Performance Qualification2
The product/process PQ is documented verification that the process and/or the total
process-related system perform as intended throughout their normal operating ranges.
In practice, this process ensures that the system (computer hardware, computer
software, controlled equipment, interfaces, operating environment, operators, equipment,
procedures, and so on), in its normal operating environment, produces acceptable quality
product and that sufficient documentary evidence exists to demonstrate this assessment.
Traditionally, the qualification documents described above contain both specification
information and instructions on testing.
2
Note: This activity is the culmination of computer systems controlling manufacturing equipment
and is out of the scope of computer systems validation. This qualification may be performed by QA
or R&D and only on manufacturing systems.
Appendix E
229
Formal Release for Use

An appropriate signature sign-off, as defined in the validation plan, must be obtained
prior to officially utilizing the system in its operating environment. Once the system is
considered validated, all applicable policies and procedures governing the operation of
the validated computer system become effective. A formal written notification should be
issued to the system users and to QA to officially inform them that the system is validated
and can be used in operational environment.
After the computer system is put into operation and all appropriate validation
documentation has been collected, as defined in the validation plan, the entire validation
documentation package should be archived to ensure its safety and security.
Validation Process Deviations
A deviation is a completed event not conforming with the validation plan or validation
procedures. It may include documentation errors, test results not meeting the acceptance
criteria, or deviations to a procedure.
When a deviation takes place, the project manager and/or tester executing the protocol
must complete a test incident (or similar) report. This report describes the nature of the
event, the corrective action(s) necessary to resolve the event, an assessment of the impact
of the event on the test case or any related documentation, software and any subsequent
test cases and/ or protocols.
The validation process deviations must be written in a procedure.
OPERATIONAL LIFE
RETIREMENT
The retirement of computer systems, which have been used in the development, clinical,
manufacture, packaging, or marketing of regulated operations, is a critical process for a
company. The purpose of the SLC retirement period is to replace or eliminate the
current computer technologies supporting a regulated operation, and to ensure the
availability of the data generated by it for conversion, migration, or retirement.
It is a requirement of Part 11 that all data produced from electronic files is in human
readable form and that true copies are available. In the case of data retirement, the
planning must include the technologies needed to support the retired records, for the
duration of the retention period required in the applicable regulations. If electronic
records which have been retired need to be accessed in the future for review, the data and
the application source code need to be readily accessible. The qualification of any
Appendix E
230
conversion processes for the data prior to archive/retirement must pay special attention to
these points.
A decommissioning plan should be written prior to the retirement of a system to
ensure continuity of service.
In some cases, a replacement system will be installed which can access the data
records from the previous system, including historical data, to enable such data to be
available in the format used by the new system. If necessary, a data conversion process
must be carried out. Where system obsolescence forces the need to convert electronic
records from one system or format to another, the conversion process must be qualified
and its integrity verified. The electronic copies of electronic record files can be
considered to be electronic record files provided that the data transfer processes have
been verified. Services such as Digital Notary expedite the verification of the conversion
process.
Following the successful conversion of the electronic records, and QA acceptance of
the validity of the conversion, the new system may be authorized for use. The results of
the validation must be fully documented and added to the validation file of the system
that is being retired.
In other cases, the electronic records will not be converted for use by the new system.
In this case, the obsolete system could be maintained as a legacy system, and utilized to
access the electronic records. This approach can be expensive and is one that might still
require the transfer of the legacy data to the new system or format at a later date, when
maintenance becomes impractical.
During product research, production, and control of FDA regulated products,
documentary evidence must be retained for a certain period, which prove the safety and
efficacy of the product. When a computer system is to be retired from active use, the data
from that system must be archived. Some regulations which cover FDA regulated
products, and which require records to be retained, are listed in Table E1.
Table E1. Regulations Which Require the

Retention of Records.
AREA SECTION
GLP
GLP
GLP
GLP
GLP
GLP
GLP
GLP
GLP
GLP
GCP
GCP
REGULATION
Organization and personnel records

FDA 58.3
Specimen and data storage facilities
FDA 58.51
Storage and retrieval of records and data FDA 58.190 & 58.195
Organization and personnel
OECD 1.1 & 1.2
Sampling and storage
OECD 6.1
Archive facilities
OECD 3.4
Characterization
OECD 6.2
SOPs
OECD 7.2
Storage and retrieval
OECD 10.1 & 10.2
Storage and retrieval with
OECD 10.2
archive facilities
Supply and handling of
ICH 5.14
investigational product(s)
Trial management, data handling,
ICH 5.5
record keeping
Appendix E
GMP
GMP
GMP
Records and documentation

Personnel records
Documentation
231
FDA 211.180
EU PIC 2
EU PIC 4
The company may wish to keep other data produced during the course of an operation,
e.g., data on products never sent for production. This would be a business decision for the
company, and is not a regulatory requirement.
In addition to the retention of operating data in electronic format, the following list
shows the supporting documentation that should be archived along with the data. The
integrity of archived data is assured when the systems on which the data was produced
have been validated, and the documentation relating to these systems has been assembled,
as part of the validation file.
This is the key documentation to be retained for a validated system due for retirement:
IQ, OQ, PQ protocols and reports (when applicable)
Project and Quality Plan
Process/Operation and requirements specification
criticality analysis
computer systetm specification
system drawings
system (hardware and software) configuration and design
source code (if available)
application software documentation.
system testing
development history and/or system development files
system manuals
user manuals
technical manuals
list of users trained to use the system and proof of the training records
technical training manuals
equipment inventory list
inventory list (hardware and software)
calibration list
log book(s)
history of changes to the system
history of audits
When all computer systems decommissioning activities have been completed, the
personnel responsible for performing the decommissioning activities may issue a memo
that indicates the following:
the reason for decommissioning the system
the date that the system will be decommissioned
the system documentation and electronic records retention period expiration date
indicating a final date for which the system documentation must be retained, following
system decommission
signatures from the appropriate personnel
Appendix E
232
signatures from the designated personnel involved in performing and verifying

completion of the decommissioned activities
COMPUTER SYSTEM INCIDENTS

The malfunction or failure of system components, incorrect documentation, or improper
operation that makes the proper use of the system impossible for an undetermined period,
characterizes some of the incidents that can affect the correct operation of a computer
system. These system incidents are also nonconformances.
In order to remedy problems quickly or to avoid system errors, appropriate procedural
controls must be applied to (a) enable early identification of an imminent breakdown, and
(b) initiate appropriate preventive action. The next section suggests a course of action to
follow to identify, analyze, and record computer system incidents, and to implement and
test the corrective action taken.
A qualified person should conduct a root cause analysis in order to determine the nature
of the malfunction or error. If it is determined that the malfunction is equipment
related, the following process may be implemented.
The personnel responsible for managing the component causing the incident must:
complete a problem description form and include an initial error analysis
record the incident using an incident log
Check any stored data and, if applicable, reconstruct the data.
If the error analysis demonstrates that a malfunction or failure of the software, or of a
hardware component has occurred, then:
using a root cause analysis, identify and specify the modification needed to fix the
malfunction
fill out the appropriate change control forms and submit these for approval. Include
the root cause analysis report
if the system modification or replacement part does not constitute an emergency
change, wait until the change control forms are approved
if the system modification or replacement part constitute an emergency change, a
qualified person may make the necessary modification(s) immediately. In this case,
after the modification have been made, the system can return to production and the
following steps implemented, as appropriate
perform a criticality and complexity assessment in order to determine the
documentation requirements for the modification and to enable the approved
system development lifecycle model to be followed
if the system modification or replacement part is not an emergency, carry out the
corrective actions identified
if a test environment is available, test the software/hardware that have been modified
before the system is returned to the operational environment
if applicable, update the system documentation and/or users manual
Appendix E
233
if the users manual was updated, conduct user training on the changes
if the analysis demonstrates incorrect documentation
update the system documentation and/or users manual
conduct user training on the revised documentation
if the analysis demonstrates improper operation
conduct user refresher training on the correct operation of the system
Release the system for use
If the analysis uncovers incorrect documentation or improper operation, the originator of
the report, the area manager and the technical manager(s) must approve the incident
report, including the recommendation made, and send it to the facilitys quality assurance
department for review and approval. The signed report should be returned to the
originator, system administrator, or system owner for implementation. On completion of
the implementation, the original report and associated documentation must be filed in the
site document retention center, and/or with the change control documentation for the
computer system.
If an observed incident constitutes an emergency, a qualified person may make
changes immediately. In this case, after the changes have been implemented, the system
can return to production. The area manager and the site quality assurance department
must be notified of the changes(s) in writing before the close of business on the day after
the incident, a formal incident report must be filed before the close of business on the day
after the incident, and all other required processes and documentation must be completed.
The processes may include (as appropriate): the review and update of the Requirements
Traceability Matrix, an update of the requirements, computer system and technical design
specification, deliverables, the development of integration tests, the modification and
application of Installation Qualification and Operational Qualification in both the test
environment and the production environment, the modification and application of
Performance Qualification, the modification of manuals and training materials, the
presentation of incremental user training, and the preparation of the Change
Implementation Report.
If incidents are managed electronically, all necessary controls must be in place for the
creation, maintenance, archiving, review, and approval of the incidents.
In both the paper-based and electronic-based environment the System Incident Report
or its equivalent must be retained according to the companys record-retention policies
after the incidents are closed (See Figure E3).
Appendix E
234
Figure E3. Sample Development

Activities and Work Products.
Appendix F
Administrative Procedures Mapped to Part
111
ADMINISTRATIVE PROCEDURES NECESSARY TO GUARD
DATA INTEGRITY, CONFIDENTIALITY, AND AVAILABILITY
Requirement
Implementation
Certification
Evaluation to certify that the computer and network are
appropriately secure.
Chain of trust partner agreement
If data is processed through a third party, the parties will be
required to enter into a chain of trust, or similar, partner
agreement.
Contingency plan (all listed implementation features
must be present)
The contingency plan enables a system emergency to be
responded to and includes:
Applications and data criticality analysis
Data backup plan
Disaster recovery plan
Plan for the emergency operating mode
Testing and revision
Formal mechanism for processing records
Documented SOPs for the routine and nonroutine receipt,
manipulation, storage, dissemination, transmission, and/or
disposal of information.
Information accesses control (all listed implementation
features must be present)
Granting users different levels of access to information.
Internal audit
An in-house review of the records of system activities (e.g.,
logins, file accesses, security incidents)
1
Applications and data

criticality
Data backup plan
Plan for the emergency
operating mode
Testing and revision
procedures
Access authorization
Access establishment
Access modification
Department of Health and Human Services, Office of the Secretary, 45 CFR Part 142 Security
and Electronic Signature Standards; Proposed Rule, Federal Register 63 (155), 4324243280,
August 12, 1998.
Appendix F
Requirement
236
Implementation
Supervision of maintenance personnel by an

authorized, knowledgeable person
Maintenance of record of access
authorizations
Operating, and in some cases, maintenance
personnel have proper access authorization
Personnel clearance procedure
Personnel security policy/procedure
System users, including maintenance
personnel, are provided with security training
Security configuration management (all Documentation
Hardware/software installation and
listed implementation features must be
maintenance review and the testing of the
present)
The organization to implement the necessary security features
measures, practices, and procedures for the System inventory
security of information systems
Security testing
Virus checking
Report procedures
Security incident procedures (all listed
implementation features must be present) Response procedures
Implement accurate and current security
incident procedures. The reporting of
security breaches, so that security violations
are reported and handled promptly.
Security management process (all listed Risk analysis
implementation features must be present) Risk management
Creating, administering, and overseeing
Sanction policy
procedures to ensure the prevention,
Security policy
detection, containment, and correction of
security breaches.
Combination locks changed
Termination procedures (all listed
implementation features must be present) Removal from access lists
Procedures for the ending of an employees Removal of user account(s)
employment or an internal/external users Return of keys, tokens, or cards that allow
access.
access
Awareness training for all personnel
Training (all listed implementation
(including management)
Security training for all staff on the
Periodic security reminders
procedures that must be followed to ensure User education concerning virus protection
the protection of information.
User education on the importance of
monitoring logins, success/failure of system
logins, and how to report discrepancies
User education in password management
Personnel security (all listed
implementation features must be present)
All personnel with access to health
information must be authorized to do so
after receiving the appropriate clearance.
Appendix F
237
PHYSICAL SAFEGUARDS NECESSARY TO GUARD DATA

INTEGRITY, CONFIDENTIALITY, AND AVAILABILITY
Requirement
Implementation
Assigned security responsibility
The assignment of responsibility for security to a specific individual or
organization.
Media controls (all listed implementation features must be present) Access control
This area is related to the procedures that govern the receipt and
Accountability
removal of hardware/software/data.
(tracking
mechanism)
Data backup
Data storage
Disposal
Physical access controls (limited access) (all listed implementation Disaster
recovery
This area is related to the procedures for limiting physical access to an Emergency
entity while ensuring that properly authorized access is allowed.
operating mode
Equipment
control (into and
out of site)
Facility security
plan
Procedures for
verifying access
authorizations
prior to physical
access
Maintenance
records
Need-to-know
procedures for
personnel access
Sign-in for
visitors and their
escorts,
if appropriate
Testing and
revision
Policy/guideline on workstation use
Procedures to maximize the security of information (e.g., logging off
before leaving a terminal unattended).
Secure workstation location
This area is related to the physical safeguards that eliminate or
minimize the possibility of unauthorized access to information.
Security awareness training
This is a requirement for all employees and contractors to understand
their security responsibilities, based on job responsibilities in the
organization, and to make security a part of their daily activities.
Appendix F
238
TECHNICAL SECURITY SERVICES NECESSARY TO GUARD

DATA INTEGRITY, CONFIDENTIALITY, AND AVAILABILITY
Requirement
Implementation
Access control Note: Emergency access procedures must be
present, and at at least one of the following three implementation
features must be present: context-based access, role-based access,
user-based access. The use of encryption is optional.
The objective is to implement an access control service that would
restrict access to resources and that would only allow access by
privileged entities.
Audit controls
Put in place audit control mechanism to record and examine system
activity.
Authorization control (at least one of the listed implementation
See access control above.
Data authentication
To provide evidence that the data possessed by the system has not
been altered or destroyed in an unauthorized manner. Examples of
how data authentication may be assured include the use of an
authentication code, or digital signature.
Entity Authentication
The following implementation features must be present:
Automatic logoff
Unique user identification
In addition, at least one of the other listed implementation features
must be present to verify that an entity is who it claims to be.
Context-based
access
Emergency
access procedures
Encryption
Role-based
access
User-based
access
Role-based
access
User-based
access
Automatic logoff
Biometrics
Password
PIN
Telephone
callback
Token
Unique user
identification
Appendix F
239
TECHNICAL SECURITY MECHANISMS NECESSARY TO

GUARD AGAINST UNAUTHORIZED ACCESS TO DATA THAT
IS TRANSMITTED OVER A COMMUNICATIONS NETWORK
Requirement
Implementation
Communications/network controls
The following implementation features must be present:
integrity controls
message authentication
If communications or networking is employed, one of the
following implementation features must be present:
access controls
encryption
In addition, if using a network, the following implementation
features must be present:
alarm
audit trail
entity authentication
event reporting
access controls
alarm
audit trail
encryption
entity authentication
event reporting
integrity controls
message
authentication
Appendix G
Sample Audit Checklist for a Closed
System
ELECTRONIC SIGNATURE AND ELECTRONIC RECORDS
AUDIT CHECKLIST FOR A CLOSED SYSTEM
General Vendor Checklist
Appendix G objective:
The purpose of the following checklist is to help to determine if a computer system
complies with the FDA Rule 21 CFR 21 Part 11 for electronic records and
electronic signatures. This audit questionnaire applies to systems that meet the
definition of a closed system as defined in Section 11.3 (b)(4) of the rule and
which do not utilize biometrics identification methods.
Section Question
Answer
Comments
Yes No N/A
11.10(b) Are data entries stored in the database tables in an

encrypted form?
11.10(b) Can data entries be printed directly from the database,
then read by a human without a key?
11.10(b) Can records be output in common formats such as
ASCII, VSAM, XLS, PDF, etc.?
11.10(e) Is every data or signature entry recorded in the database
11.50 with a date and time stamp?
11.10(e) Is an entry that has been made and then later corrected
retained in the system?
11.10(e) Is a prior entry that has been corrected identified as
superseded?
Section Question
11.10(e) If an entry is made by the operator, but is not saved due
to a warning about its potential incorrectness retained in
the system?
11.10(e) Does the system include in its audit trail operator
actions that are not based upon a data entry or signature,
(e.g., when an operator triggers an action, responds to a
message, etc.?
11.10(d) Does the system limit who may access the system
functions?
Answer
Comments
Yes No N/A
Appendix G
242
11.10(g) Does the system restrict individuals to specific

documents?
11.10(g) Does the system restrict individuals to specific
11.10(h) workstations, functions, or documents?
11.10(f) Does the system have the capability to control the
sequence in which users may view screens?
11.10(f) Does the system have the capability to detect when an
entry occurs outside of the normal sequence?
11.10(h) Does the system record the location (node) of the
workstation where each entry was made?
11.50(b) Does the database table for each entry show the
unencrypted name of the entry maker?
11.50(a) Does the system store the meaning or description of
each signature
11.50(b) Are the details of the signing (e.g., signer, date, time,
meaning) included on every display and printout?
11.70
Does the system have an absolute method of linking
each signature to its corresponding entry?
11.200(a) Does each signature entry have two separate methods
for identifying who is making that signature, and are
both of these supplied by the operator every time he
signs his name?
Section Question
11.10(c) Can inactive records that are still within the legal
retention period be readily retrieved, completely intact,
with their full audit trail?
11.10(c) Does the system have a security scheme to protect
database(s) from accidental or deliberate damage?
11.10(k) Is there a written procedure in place for the control,
change, distribution, and destruction of system
documentation?
11.10(k) Are written records kept of the copies and possession
of copies for all system documentation?
11.10(k) Does every system documentation change have
approvals and a document change history?
11.100 Is there any possibility that two people could share the
11.300(a) same signature components at any time in the past,
present, or future?
11.100 Is there any possibility that two or more people could
share the same name (e.g., for a database table)?
1.100(b) Is there a written form for the assignment of electronic
signatures?
1.100(b) Does the signature assignment procedure require
positive identification of the person to whom the
signature codes are given?
11.200 Is there any way that one person could learn and/or use
Answer
Comments
Yes No N/A
Appendix G
243
the signature codes of another person?

11.300(d) Does the system have a method to detect the
unauthorized use of signature codes?
11.10(a) Is there a current system validation for all system
software and hardware components?
Section Question
11.10(a)
11.300(b)
11.300(b)
11.300(c)
11.300(c)
11.300(e)
11.300(e)
11.10(i)
11.10(i)
11.10(i)
11.100(c)
11.100(c)
Does the system have a method to detect the

alteration of any record, even if the alteration was
done directly to the database?
Are signature codes periodically changed or
verified?
Are there records in existence that confirm that
signature codes are periodically changed or verified?
Is there a written procedure in place for the
replacement of any objects (tokens) that are used for
signature purposes when they become damaged or
lost?
Are there records in existence that confirm that the
procedure for the replacement of objects (tokens) is
being followed?
Is there a procedure in place for the periodic testing
of any objects (tokens) that are used for signature
purposes?
Are there records in existence that confirm that the
procedure for the periodic testing of objects (token)
is being followed?
Are there records or statements of qualifications for
all of the persons who developed, installed, and/or
validated the system?
Is there a user-training program with course
materials for all categories of system user?
Are there training records/assessments of the user
training received for every system user?
Has the company/location notified the FDA that it is
using electronic signatures?
Has the company determined the method that it will
use to supply additional testimony as to the validity
of a specific signature?
Answer
Comments
Yes No N/A
Appendix H
Computer Systems Regulatory
Requirements
This appendix discusses the main FDA regulatory requirements and guidelines applicable
to computer systems performing regulated operations and their interpretation. The
following regulations are covered:
Guidelines on General Principles of Process Validation, May 1987
Human and veterinary products GMP Regulations (21 CFR 210 and 211)
Nonclinical laboratory GLP Regulations (21 CFR 58)
Medical Devices GMP Regulations (21 CFR 820)
Food (21 CFR 110 and others)
GUIDELINES ON GENERAL PRINCIPLES OF PROCESS

VALIDATION
This guideline, developed in 1987, outlines the general principles considered by the FDA
as acceptable elements of process validation for the preparation of human and animal
drug products, and for medical devices. Our analysis of the elements impacting
computer systems will concentrate on Section VIII Elements of Process Validation. The
connection between the Process Validation guideline and computer systems performing
regulated operations is established in CPG 7132a.11.
VIII (A)PROSPECTIVE VALIDATION

Prospective validation includes considerations that need to be made before a firm
introduces an entirely new product, or when there is a change in the manufacturing
process that may affect the products characteristics, such as uniformity and identity.
ANALYSIS
1 All computer systems performing regulated operations should be validated; the depth
and scope of validation depends on the type of software and the complexity and
criticality of the application.
2 Validation should be performed to verify the identifiable performance requirements of
an application.
Appendix H
245
3 Before a system is brought into use, it should be thoroughly tested and confirmed as
being capable of achieving the desired results and performance.
4 All new computer systems must be prospectively validated before they are brought into
operational use.
VIII (A)(1)EQUIPMENT AND PROCESS

The equipment and process(es) should be designed and/or selected so that product
specifications are consistently achieved. This should be done with the participation of all
appropriate groups concerned with assuring product quality, e.g., engineering design,
production operations, and quality assurance personnel.
ANALYSIS
1 Attention should be paid to the siting of computer hardware in suitable conditions
where extraneous factors cannot interfere with the system.
2 Based on the identification of operational functions, design and proper computer
performance, relevant computer hardware technologies can be selected. Depending on
the available technology and its cost, automated functions can be assigned to the
computer hardware and/or software.
3 A set of design documentation, including as-built drawings, should be maintained for
computers, infrastructure, and instrumentation.
4 Integration of hardware and the development and testing of hardware and software
should be performed under a quality assurance system. For computer technology
suppliers and developers, the quality assurance system must be agreed upon and
defined in a contract between the supplier and customer.
VIII (A)(1)(A)EQUIPMENT: INSTALLATION

QUALIFICATION
Installation qualification studies establish a degree of confidence that the process
equipment and ancillary systems are capable of consistently operating within established
limits and tolerances. After process equipment is designed or selected, it should be
evaluated and tested to verify that it is capable of operating satisfactorily within the
operating limits required by the process. This phase of validation includes examination of
equipment design; determination of calibration, maintenance, and adjustment
requirements; and the identification of critical equipment features that could affect the
process and product. Information obtained from these studies should be used to establish
written procedures covering equipment calibration, maintenance, monitoring, and control.
In assessing the suitability of a given piece of equipment, it is usually insufficient to
rely solely upon the representations of the equipment supplier, or upon experience in
producing some other product. Sound theoretical and practical engineering principles and
considerations are the first steps in the assessment.
Appendix H
246
It is important that equipment qualifications simulate actual production conditions,

including those that are worst-case situations. Tests and challenges should be repeated a
sufficient number of times to assure reliable and meaningful results. All acceptance
criteria must be met during the test or challenge. If any test or challenge shows that the
equipment does not perform within its specifications, an evaluation should be performed
to identify the cause of the failure. Corrections should be made and additional test runs
performed, as needed, to verify that the equipment performs within specifications.
The observed variability of the equipment between and within runs can be used as a
basis for determining the total number of trials selected for the subsequent performance
qualification studies of the process. Once the equipment configuration and performance
characteristics are established and qualified, they should be documented. The installation
qualification should include a review of pertinent maintenance procedures, repair parts
lists, and calibration methods for each piece of equipment. The objective is to assure that
all repairs can be performed in a manner that will not affect the characteristics of material
processed after the repair.
In addition, special post-repair cleaning and calibration requirements should be
developed to prevent inadvertent manufacture of a nonconforming product. Planning
during the qualification phase can prevent confusion during emergency repairs that could
lead to the use of the wrong replacement part.
ANALYSIS
1 Assurance of computer systems reliability is achieved by the execution of a quality plan
and by testing during the software development process. This involves unit code
testing and integration testing in accord with the principles specified in ISO 90003
and IEEE 1298 (PIC/S).
2 Appropriate IQ, OQ and PQ tests should demonstrate the suitability of computer
hardware and software to perform their assigned tasks and to ensure their proper
performance in an operational environment. While IQ, OQ, and PQ terminology has
served its purpose well, and is one of many legitimate ways to organize dynamic
testing in the operational environment for the FDA-regulated industries, this
terminology may not be well understood among many software professionals.
However, organizations performing regulated operations must be aware of these
differences in terminology as they request and provide information regarding
computer systems.
3 The software is a critical component of a computer system. The user should take all
reasonable steps to ensure that it has been produced under a quality assurance system.
4 Written procedures should be available to:
operate the system under normal and abnormal situations
detect and record errors, and enable corrections to be made
restore a system to an operational condition including data recovery
authorize, carry out, and record changes
enable, when appropriate, electronic signatures
Refer to Chapter 12 and Appendix F.
Appendix H
247
VIII (A)(2)SYSTEM TO ASSURE TIMELY REVALIDATION

There should be a quality assurance system in place that requires revalidation whenever
there are changes in packaging, formulation, equipment, or processes that could impact
on product effectiveness or product characteristics, and whenever there are changes in
product characteristics. Furthermore, when a change of raw materials supplier occurs, the
manufacturer should consider subtle, potentially adverse differences in the raw material
characteristics. A determination of adverse differences in raw material indicates a need to
revalidate the process.
One way of detecting the kind of changes that should initiate revalidation is the use of
tests and methods of analysis capable of measuring characteristics that may vary. Such
tests and methods usually yield specific results going beyond the mere pass/fail basis,
thereby detecting variations within product and process specifications and allowing
determination of whether a process is slipping out of control.
The quality assurance procedures should establish the circumstances under which
revalidation is required. These may be based upon equipment, process, and product
performance observed during the initial validation challenge studies. It is desirable to
designate individuals who have the responsibility to review product, process, equipment,
and personnel changes to determine if and when revalidation is warranted.
The extent of revalidation will depend upon the nature of the changes and how they
impact upon different aspects of previously validated production. It may not be necessary
to revalidate a process from scratch merely because a given circumstance has changed.
However, it is important to carefully assess the nature of the change to determine
potential ripple effects, and what needs to be considered as part of revalidation.
ANALYSIS
1 All changes made to the computer system should be formally authorized, documented,
and tested. Records should be kept of all changes including modifications and
enhancements made to the hardware, software and any other critical component of the
system to demonstrate that the final system has been maintained in a validated state.
2 When appropriate, regression testing (refer to Chapter 9) must be performed. The depth
and scope of revalidation depends on the type of software and the complexity, and
criticality of the application.
3 All revalidations must be prospective, before they are brought into operational use.
VIII (A)(3)DOCUMENTATION
It is essential that the validation program is documented and that the documentation is
properly maintained. Approval and release of the process for use in routine
manufacturing should be based upon a review of all the validation documentation,
including data from the equipment qualification, process performance qualification and
product/package testing to ensure compatibility with the process.
Appendix H
248
For routine production, it is important to adequately record process details (e.g., time,
temperature, equipment used) and to record all changes that have occurred. A
maintenance log can be useful in performing failure investigations concerning a specific
manufacturing lot. Validation data (along with specific test data) may also determine
expected variance in product or equipment characteristics.
ANALYSIS
1 Refer to Chapter 2, Key Validation Elements.
GOOD MANUFACTURING PRACTICE REGULATIONS

RELATED TO COMPUTER SYSTEMS VALIDATION
The GMP regulations have been in existence since 1963. The current regulations are from
1978 (43 FR 45077), and amended in 1995 (60 FR 4091). The proposed GMP regulation
from 1996, includes Section 211.220(a) that would require the validation of computer
systems performing manufacturing-regulated operations. In addition to section
211.220(a), several other sections apply to computer systems and to responsible
individuals.
SUB-PART BORGANIZATION AND PERSONNEL

211.25 Personnel Qualifications
(a) Each person engaged in the manufacture, processing, packing, or holding of a drug
product shall have education, training, and experience or any combination thereof, to
perform the assigned functions. Training shall be in the particular operations that the
employees perform and in current good manufacturing practice.
(b) Each person responsible for supervising the manufacture, processing, packing, or
holding of a drug product shall have the education, training, and experience or any
combination thereof, to perform assigned functions in such a manner as to provide
assurance that the drug product has the safety, identity, strength, quality, and purity
that it purports or is represented to possess.
ANALYSIS
1 Appropriate training must be provided for personnel involved with computer systems
and must also include training in GMPS.
2 The documentation of qualifications including the education, training, experience, or
any combination thereof must be available for developers, users, and support
personnel.
Appendix H
249
SUB-PART CBUILDING AND FACILITY

211.42 Design and Construction Features
(a) Any building or buildings used in the manufacture, processing, packing, or holding of
a drug product shall be of suitable size, construction and location, to facilitate
cleaning, maintenance, and proper operations.
ANALYSIS
The area allocated to the computer system must be suitable for the proper operation of the
system.
SUB-PART DEQUIPMENT
211.63 Equipment Construction
Equipment used in the manufacture, processing, packing, or holding of a drug product
shall be of appropriate design, adequate size, and suitably located to facilitate operations
for its cleaning and maintenance.
ANALYSIS
1 Computer hardware must be properly specified to meet the requirements for its
intended use, and the amount of data it must handle.
2 Environmental controls, electrical requirements, electromagnetic noise control, and
others should be considered when determining a location for the hardware.
3 The location of the hardware must allow access for maintenance, as required.
211.68 Automatic, Mechanical, and Electronic Equipment
(a) Automatic, mechanical, or electronic equipment or other types of equipment,
including computers, or related systems that will perform a function satisfactorily,
may be used in the manufacture, processing, packing, and holding of a drug product. If
such equipment is so used, it shall be routinely calibrated, inspected, or checked
according to a written program designed to assure proper performance. Written
records of those calibration checks and inspections shall be maintained.
(b) Appropriate controls shall be exercised over computer or related systems to ensure
that changes in master production and control records or other records are instituted
only by authorized personnel. Input to and output from the computer or related system
of formulae or other records or data shall be checked for accuracy The degree and
frequency of input/output verification shall be based on the complexity and reliability
Appendix H
250
of the computer or related system. A backup file of data entered into the computer or
related system shall be maintained except where certain data, such as calculations
performed in connection with laboratory analysis, are eliminated by computerization
or other automated processes. In such instances a written record of the program shall
be maintained along with data establishing proper performance. Hard copy or
alternative systems, such as duplicates, tapes, or microfilm, designed to assure that
backup data are exact and complete and that it is secure from alteration, inadvertent
erasures, or loss shall be maintained.
ANALYSIS
1 There must be a written program detailing the maintenance of the computer system
(i.e., hardware maintenance manual). The maintenance of the computer, including
periodic scheduled maintenance and breakdown maintenance, must be documented.
2 There must be a system to control changes to both the hardware and the software.
Changes must only be made by authorized individuals following an appropriate review
and approval of the change.
3 There must be documented checks of the system inputs and outputs for accuracy (refer
to CPG 7132a.07, I/O Checking). According to this CPG, computers I/Os are to be
tested for data accuracy as part of the computer system qualification and, after the
qualification, as part of the computer systems ongoing verification program. This
CPG and the proposed 1996 GMP changes in Part 211.68(b) are considered to be an
implicit requirement to validate all computer systems covered under the GMPS.
Proposed 211.220(a) and Part 11.10(a) explicitly require validation of computer
systems.
4 Computer files/data must be backed up either by hard copy, microfilm, or magnetic
media. Computer files/data must be maintained electronically.
5 There must be validation data and a written record of the programs used when backup
data is eliminated by computerization.
SUB-PART FPRODUCTION AND PROCESS CONTROLS

211.100Written Procedures: Deviations
(a) There shall be written procedures for production and process control designed to
assure that the drug products have the identity, strength, quality, and purity they
purport or are represented to possess. These written procedures, including any charges,
shall bereviewed and approved by the quality control unit.
ANALYSIS
1 There must be written procedures covering the operation of computer systems.
Appendix H
251
2 The Quality Assurance unit must review and approve procedural controls for computer
systems.
SUB-PART JRECORDS AND REPORTS

211.180 General Requirements
(a) Any production, control, or distribution recordspecifically associated with a batch
of drug product shall be retained for at least one year after the expiration date of the
batch.
ANALYSIS
1 Electronic records must be maintained electronically. In the case of electronic batch
records, they must be retained, as a minimum, for at least one year after the expiration
date of the batch.
2 The retention periods for supporting documentation such as maintenance records,
change control records and project documentation are as follows:
(a) obsolete system documentationone year after the expiration date of the last batch
supported by the legacy function or system;
(b) retired system documentationone year after the expiration date of the last batch
supported by the retired functionality/ system; current documentationretain
permanently until (a) or (b) above apply.
GOOD LABORATORY PRACTICE REGULATIONS RELATED

TO COMPUTER SYSTEMS VALIDATION
The GLP regulations have been in existence since 1978. From 1984 the FDA began
applying these regulations to computer systems, their development, use, and support. The
FDA has stated that hardware is considered to be equipment and software is considered
to be records under the regulations. This means that computer system development
documentation and procedures are now subject to FDA inspection.
The following sections provide an interpretation of how the GLPs can be applied to
computer systems and responsible individuals (note that the word systems may be
substituted for studies in the regulations):
Part 58Good Laboratory Practice For Nonclinical Laboratory
Studies
Computer systems are often used in nonclinical laboratory studies to record, analyze,
store, and summarize data. In this context, the computer systems are subject Good
Laboratory Practice (GLP) regulations.
Appendix H
252
Compliance with these regulations is intended to ensure the quality and integrity of the
safety data that is filed.
SUB-PART BORGANIZATION AND PERSONNEL

58.29 Personnel
(a) Each individual engaged in the conduct ofa nonclinical laboratory study shall have
education, training, and experience, or a combination thereof, to enable that individual
to perform the assigned function.
(b) Each testing facility shall maintain a current summary of training and experience and
job description for each individual engaged inthe conduct of a laboratory study.
ANALYSIS
1 Appropriate training must be provided for personnel involved in the computer system
development, use, and support.
2 Maintain current job descriptions, training records, and a summary of the education
received and qualification gained (e.g., a Curriculum Vitae).
58.31 Testing Facility Management
For each nonclinical laboratory study, testing facility management shall(c) assure that
there is a quality assurance unit(e) ensure that personnel, resources, equipment,
materials, and methodologies are available as scheduled, (f) assure that personnel clearly
understand the functions they are to perform, (g) assure that any deviations from these
regulations reported by the quality assurance unit are communicatedand corrective
actions are taken and documented.
ANALYSIS
1 Management is responsible for establishing and maintaining an effective GLP
compliance program.
2 Management assures that:
(a) personnel, facilities, equipment, materials, and methodologies are available as
scheduled.
(b) responsible personnel are available as scheduled and they clearly understand and
carry out the functions that they are to perform.
(c) Quality Assurance is established, and all deviations are reported, corrected, and
documented.
Appendix H
253
58.35 Quality Assurance Unit

(a) Testing facility shall have a quality assurance unitwho shall be responsible for
monitoring each study to assure management that the facilities, equipment, personnel,
methods, practices, records and controls are in conformance with the regulationsthe
quality assurance unit shall be entirely separate from and independent of the personnel
engaged in the direction and conduct of that study.
(b) The quality assurance unit shall:
Inspect each phase of a nonclinical laboratory study periodically and maintain
written and properly signed recordsshowing the date of the inspection, the
study inspected, the phase of the study inspected, the person performing the
inspection, findings and problems, action recommended and taken to resolve
existing, problems, and any scheduled date for reinspectionsany significant
problems, which are likely to affect study integrity found during the course of an
inspection, shall be brought to the attention ofmanagement immediately.
Periodically submit to managementwritten status reportsnoting any problems
and the corrective actions taken.
Determine that no deviations from approved protocols or standard operations
procedures were made without proper authorization and documentation.
Prepare and sign a statementwhich shall specify the dates inspections were
made and the findings reported to management.
(c) The responsibilities and procedures applicable to the quality assurance unit the
records maintainedand the method of indexing such records shall be in writing and
shall be maintained. These items including inspection dates, the study inspected, and
the name of the individual performing the inspection shall be made available for
inspection to authorized employees of the Food and Drug Administration.
(d) A designed representative of the Food and Drug Administration shall have access to
the written procedures established for the inspection and may request testing the
facility management to certify that inspections are being implemented, performed,
documented, and followed up.
(e) All records maintained by the quality assurance unit shall be kept in one location at
the testing facility.
ANALYSIS
Quality Assurance must be established for the following:
1 Monitoring systems to assure management that the facilities, equipment, personnel,
methods, practices, records and controls are in conformance with the regulations.
2 Exist entirely separate from and independent of the personnel engaged in developing
and use of laboratory computer systems.
3 Inspect systems at intervals adequate to ensure the integrity of the system and its data.
4 Determine that procedural controls have been written, and that no deviations from them
were made without proper authorization and documentation.
Appendix H
254
5 Retain documented and properly signed records of each inspection, indicating the date,
what was inspected, responsible individual, findings/problems, resolution, and the
scheduled date for reinspection.
6 Report to management immediately any significant problems likely to affect study
integrity.
7 Submit periodic written reports of findings to management, and include any deviations
from the regulations and the corrective action taken.
8 Maintain QA procedural controls for responsibilities and procedures, and make them
available for inspection by the FDA.
SUB-PART CFACILITIES
58.41 General
Each testing facility shall be of suitable size, construction, and location to facilitate the
proper conduct of nonclinical laboratory studies.
ANALYSIS
Each area defined for housing the computer system must be suitable for the proper
operation of the system.
SUB-PART DEQUIPMENT
58.61 Equipment Design
electronic equipment used in the generation, measurement, or assessment of data and
equipment used for facility environmental control shall be of appropriate design and
adequate capacity to function according to the protocol and shall be suitably located for
operation, inspection, cleaning, and maintenance.
58.63 Maintenance and Calibration of Equipment
(a) Equipment shall be adequately inspected, cleaned and maintained. Equipment used
for the generation, measurement, or assessment of data shall be adequately tested,
calibrated, and/or standardized.
(b) The written standard operating procedures...shall set forth in sufficient detail the
methods, material, and schedules to be used in the routine inspection, cleaning,
maintenance, testing, calibration and/or standardization of equipment, and shall
specify remedial action to be taken in the event of failure or malfunction of equipment.
The written standard operating procedures shall designate a person for the
Appendix H
255
performance of each operation, and copies of the standard operating procedures shall
be made available to laboratory personnel.
(c) Written records shall be maintained for all inspection, maintenance, testing,
calibrating and/or standardizing operations. These records, containing the date of the
operation, shall describe whether the maintenance operations were routine and
followed. Written records shall be kept of nonroutine repairs performed on equipment
as a result of failure and malfunction. Such records shall document the nature of the
defect, how and when the defect was discovered, and any remedial action taken in
response to the defect.
ANALYSIS
Computer systems are composed of electronic equipment used in the generation,
measurement or assessment of data.
1 Computer systems must be appropriately designed and have adequate capacity to
function according to their specifications, and must be suitably located for evaluation
and maintenance.
2 Documented development methodologies and programming standards are required.
3 Computer systems must be adequately inspected, tested, calibrated, and maintained.
4 Documented records of routine maintenance, testing, and calibration are required.
5 Procedural controls shall define how, when, and who is responsible for the systems
inspections, testing, and maintenance (routine and nonroutine).
6 Documentation should be retained for nonroutine repairs as a result of failure and
malfunction and should include the problem, how and when it was found and the
remedial action taken.
SUB-PART ETESTING FACILITIES

58.81 Standard Operating Procedures
(a) A testing facility shall have standard operating procedures in writing setting forth
non-clinical study methods that management is satisfied are adequate to insure the
quality and integrity of the data generated in the course of a study All
deviationsfrom standard operating procedures shall be authorizedand shall be
documented.
(d) A historical file of standard operating procedures, and all revisions thereof, including
the dates of such revisions, shall be maintained.
ANALYSIS
1 Provide documented procedural controls.
Appendix H
256
2 Deviations in established procedural controls need to be properly authorized and

documented.
3 Historical files of procedural controls and all revisions are to be maintained.
Note: Procedural controls for computer applications may include: development, use, and
support of systems, data handling, storage and retrieval, maintenance, repair, calibration
of equipment, management and personnel responsibilities, record retention, training.
SUB-PART GPROTOCOL FOR AND CONDUCT OF A

NONCLINICAL LABORATORY STUDY
58.130 Conduct of a Nonclinical Laboratory Study
(e) All data generated during the conduct of a nonclinical laboratory studyshall be
dated on the day of entrythe individuals responsible for direct data input shall be
identified at the time of data input. Any change in computer entries shall be made so
as not to obscure the original entry, shall indicate the reason for a change and shall be
dated and the responsibleindividuals shall be identified.
ANALYSIS
Data must be identified (i.e., via an audit trail) with the following:
1 The individuals responsible for direct data entry.
2 The date and time of data input.
3 Documentation for all data corrections and including the original entry, reason for the
change, date and responsible individual.
Note: Data may include computer printouts, magnetic media, and recorded data from
automated instruments that are the result of the original observations, and activities of a
non-clinical laboratory study necessary for the reconstruction and evaluation of the study
report (electronic copies of electronic record files can be considered electronic record
files provided that the data transfer processes have been verified).
SUB-PART JRECORDS AND REPORTS

58.190 Storage and Retrieval of Records and Data
All raw data, documentationgenerated as a result of a nonclinical laboratory study shall
be retained.
There shall be archives for orderly storage and expedient retrieval of all raw data,
documentationConditions of storage shall minimize deterioration of the documentsin
accordance with the requirements for the time period of their retention.
(c) Only authorized personnel shall enter the archives.
Appendix H
257
ANALYSIS
1 Raw data, system records, and documentation must be retained according to the periods
stated in the regulations.
2 The data archive should be in a secure, limited access area and designed to preserve the
data, records and documents.
3 Data, system records, and documentation should be archived in a way that permits
expedient retrieval.
Note: Raw data, system records, and documentation may be retained as hard copy, on
magnetic storage media, or microforms (microfiche).
58.195 Retention of Records
(b) documentation, records, raw datashall be retained for which ever of the
following periods is the shortest:
(1) A period of at least 2 years following the date on which an application for research
or marketing permit, in support of which the results of the nonclinical laboratory
study were submitted, is approved by the Food and Drug Administration.
(2) A period of at least 5 years following the date on which the results of the
nonclinical laboratory study are submitted to the FDA in support of an application
for a research or marketing permit.
(c) Therecords of quality assurance inspectionsshall be maintained by the quality
assurance unit as an easily accessible system of records for the period of time
specified in this section.
(d) Summaries of training and experience and job descriptionsmay be retainedalong
with all otheremployment records for the length of time specified inthis section.
(f) Records and reports of the maintenance and calibration and inspection of
equipmentshall be retained for the length of the time specified in paragraph (b).
ANALYSIS
1 Support documentation such as maintenance records, change control records, and
project documentation should be retained as required by the GLPs. Refer to 58.195.
This requirement is applicable to paper-based and electronic-based records.
In addition, Quality Assurance records, summaries of training received, and job
descriptions must be kept in permanent storage.
DEVICE GMP REGULATIONS (21 CFR 820) RELATED TO

COMPUTER SYSTEMS VALIDATION
The device GMP regulations were published in 1978 (43FR 31508) and subsequently
revised in 1979 (44FR 75628), 1988 (53FR 11253), 1990 (55FR 11169) and 1996 (61FR
Appendix H
258
52601). These revised regulations require the validation of computer systems. Systems
covered by device GMP regulations include any system that directly or indirectly impacts
the safety, effectiveness, or quality of materials, components, or the finished device.
SUB-PART AGENERAL PROVISIONS

820.5 Quality System
(a) Each manufacturer shall establish and maintain a quality system that ensures that the
requirements of this part are met, and that devices produced are safe, effective, and
otherwise fit for their intended uses.
ANALYSIS
1 There must be a quality system in place appropriate to the specific device
manufactured.
2 The quality system must include provisions for assuring the validation of all computer
systems used in the manufacture of, or contained within, the finished device.
SUB-PART BQUALITY SYSTEM REQUIREMENTS

820.20 Management Responsibility
(a) Quality policy. Management with executive responsibility shall establish its policy
and objectives for, and commitment to, quality. Management with executive
responsibility shall ensure that the quality policy is understood, implemented, and
maintained at all levels of the organization.
(b) Organization. Each manufacturer shall establish and maintain an adequate
organizational structure to ensure that devices are designed and produced in
accordance with the requirements of this part.
(1) Responsibility and authority. Each manufacturer shall establish the appropriate
responsibility, authority, and interrelation of all personnel who manage, perform
and assess work affecting quality, and provide the independence and authority
necessary to perform these tasks.
(2) Resources. Each manufacturer shall provide adequate resources, including the
assignment of trained personnel, for management, performance of work, and
assessment activities, including internal quality audits, to meet the requirements of
this part.
(3) Management representative. Management with executive responsibility shall
appoint, and document such appointment of, a member of management who,
irrespective of other responsibilities, shall have established authority over and
responsibility for: (i) Ensuring that quality system requirements are effectively
Appendix H
259
established and effectively maintained in accordance with this part; and (ii)
Reporting on the performance of the quality system to management with executive
responsibility for review.
(c) Management review. Management with executive responsibility shall review the
suitability and effectiveness of the quality system at defined intervals and with
sufficient frequency according to established procedures to ensure that the quality
system satisfies the requirements of this part and the manufacturers established
quality policy and objectives. The dates and results of quality system reviews shall
be documented.
(d) Quality planning. Each manufacturer shall establish a quality plan that defines the
quality practices, resources and activities relevant to devices that are designed and
manufactured. The manufacturer shall establish how the requirements for quality
will be met.
(e) Quality system procedures. Each manufacturer shall establish quality system
procedures and instructions. An outline of the structure of the documentation used
in the quality system shall be established where appropriate.
ANALYSIS
1 The quality management system should be the responsibility of a designated
individual(s) who does not have direct responsibility for the performance of a
manufacturing operation.
2 The designated individual(s) responsible for the quality management system must have
procedures that ensure the proper review of computer system validation documents.
This function should also include the approval or rejection of all computer
hardware/software relating to the device.
3 The individual(s) responsible for the quality management system must also ensure that
adequate procedures are in place for identifying, recommending or providing solutions
for quality assurance problems relating to computer systems, and for verifying of the
implementation of such solutions.
4 There must be planned and periodic audits of the quality management system by trained
personnel who do not have direct responsibility for the area being audited. The results
of these audits must be documented in written reports that are submitted to the
appropriate management.
820.25 Personnel
(a) General. Each manufacturer shall have sufficient personnel with the necessary
education, background, training, and experience to assure that all activities required by
this part are correctly performed.
(b) Training. Each manufacturer shall establish procedures for identifying training needs
and to ensure that all personnel are trained to adequately perform their assigned
responsibilities.
Appendix H
260
ANALYSIS
1 Appropriate training must be provided for personnel involved with computer systems.
2 Quality assurance personnel must have, or be given specific training in, the auditing of
SLC processes and the identification of problems relating to the validation of
computer systems.
SUB-PART GPRODUCTION AND PROCESS CONTROLS

820.70 Production and Process Controls
(c) Environmental controls. Where environmental conditions could reasonably be
expected to have an adverse effect on product quality, the manufacturer shall establish
and maintain procedures to adequately control these environmental conditions.
Environmental control system(s) shall be periodically inspected to verify that the
system, including necessary equipment, is adequate and functioning properly. These
activities shall be documented and reviewed.
ANALYSIS
1 If environmental controls are required to ensure the proper operation of a computer
system, they must be periodically inspected and these inspections must be
documented.
2 There must be documented periodic inspections (i.e., audits) to ensure adherence to the
hardware maintenance schedule.
CLEANING AND SANITATION

820.70(d) Personnel
Each manufacturer shall establish and maintain requirements for the health,
cleanliness,personal practices, and clothing of personnel if contact between such
personnel and product orenvironment could reasonably be expected to have an adverse
effect on product quality. Themanufacturer shall ensure that maintenance and other
personnel who are required to worktemporarily under special environmental conditions
are appropriately trained or supervised bya trained individual.
Appendix H
261
820.70(e) Contamination Control

Each manufacturer shall establish and maintain procedures to prevent contamination of
equipment or product by substances that could reasonably be expected to have an adverse
effect on product quality.
ANALYSIS
1 Personnel activities such as eating, drinking or smoking, which could have an adverse
effect on a computer system, must be limited to a designated area away from the
computer system. This must be specified in a written procedure.
2 There must be written procedures designed to prevent the contamination of the
computer system.
820.70(g) Equipment
Each manufacturer shall ensure that all equipment used in the manufacturing process
meets specified requirements and is appropriately designed, constructed, placed, and
installed to facilitate maintenance, adjustment, cleaning, and use.
(1) Maintenance schedule. Each manufacturer shall establish and maintain schedules for
the adjustment, cleaning, and other maintenance of equipment to ensure that
manufacturing specifications are met. Maintenance activities, including the date and
individual(s) performing the maintenance activities, shall be documented.
(2) Inspection. Each manufacturer shall conduct periodic inspections in accordance with
established procedures to ensure adherence to applicable equipment maintenance
schedules. The inspections, including the date and individuals conducting the
inspections, shall be documented.
(3) Adjustment. Each manufacturer shall ensure that any inherent limitations or allowable
tolerances are visibly posted on or near equipment requiring periodic adjustments or
are readily available to personnel performing these adjustments.
ANALYSIS
1 Computer hardware/software must be properly specified to meet the requirements for
the use for which it is intended.
2 Environmental controls, electrical requirements, control of electromagnetic noise,
etc., should be considered.
3 The location of the hardware must allow access for cleaning and maintenance.
4 There must be a written schedule for the maintenance of the computer hardware. This
schedule should be readily available. A written record of all maintenance performed
must be maintained.
Appendix H
262
820.70(i) Automated processes

When computers or automated data processing systems are used as part of production or
the quality system, the manufacturer shall validate computer software for its intended use
according to an established protocol. All software changes shall be validated before
approval and issuance. These validation activities and results shall be documented.
ANALYSIS
1 There must be written procedures for the validation of computer systems.
2 Validation of computer systems must be performed by qualified individuals and must
be verified.
3 There must be documented checks of input and outputs for accuracy
4 Validation records must be maintained.
5 The failure of computer hardware and/or software within a finished device to meet its
specifications must be investigated. The investigation must be documented and must
include conclusions, corrective actions, and follow-up reviews.
6 There must be documented checks for programming errors.
820.72(b) Calibration
Calibration procedures shall include specific directions and limits for accuracy and
precision. When accuracy and precision limits are not met, there shall be provisions for
remedial action to reestablish the limits and to evaluate whether there was any adverse
effect on the device's quality. These activities shall be documented.
(1) Calibration standards. Calibration standards used for inspection, measuring and test
equipment shall be traceable to national or international standards. If national or
international standards are not practical or available, the manufacturer shall use an
independent reproducible standard. If no applicable standard exists, the manufacturer
shall establish and maintain an in-house standard.
(2) Calibration records. The equipment identification, calibration dates, the individual
performing each calibration and the next calibration date shall be documented. These
records shall be displayed on or near each piece of equipment or shall be readily
available to the personnel using such equipment and to the individuals responsible for
calibrating the equipment.
ANALYSIS
1 All inspection, measuring or test equipment, whether mechanical, electrical, automatic,
or mechanical in operation, which performs a quality assurance function must be
validated for its intended use. This equipment must also be routinely calibrated
according to a written procedure and the records that document these activities must
be maintained.
Appendix H
263
2 For all software used in inspection, adequate and documented testing must validate the
measuring and test equipment.
3 All hardware and/or software changes must be made by a designated individual(s)
using a formal change control procedure.
4Calibration procedures must include specific instructions and limits for accuracy and
precision. They must also include procedures for remedial action if the limits are not
met.
5Personnel performing the calibrations must have the necessary education, training,
experience, or background to perform the task. CVs should be kept on file.
6The name of the calibrator, calibration date, and the date of the next calibration must be
recorded. This information must be maintained by a designated individual(s) and must
be readily retrievable for each piece of hardware requiring calibration.
SUB-PART HACCEPTANCE ACTIVITIES

820.80(b) Receiving Inspection and Testing
Each manufacturer shall establish and maintain procedures for acceptance of incoming
product. Incoming product shall be inspected, tested, or otherwise verified as conforming
to specified requirements. Acceptance or rejection shall be documented.
ANALYSIS
1 There must be a written procedure for the acceptance/rejection of computer hardware/
software that is included in the finished device, and a designated individual(s) must
carry out this process.
820.80 (d) Final Inspection and Testing
Each manufacturer shall establish and maintain procedures for finished device acceptance
to ensure that each production run, lot, or batch of finished devices meets acceptance
criteria. Finished devices shall be held in quarantine or otherwise adequately controlled
until released. Finished devices shall not be released for distribution until:
(1) the activities required in the DMR are completed
(2) the associated data and documentation is reviewed
(3) the release is authorized by the signature of a designated individual(s), and
(4) the authorization is dated.
ANALYSIS
1 There must be written procedures for the evaluation of the finished device to assure that
all computer hardware/software specifications have been met.
Appendix H
264
SUB-PART MRECORDS
820.180 General Requirements
(b) Records retention periodAll records required by this part shall be retained for a
period of time equivalent to the design and expected life of the device, but in no case
less than 2 years from the date of release for commercial distribution by the
manufacturer.
ANALYSIS
1 Support documentation such as maintenance records, change control records, and
project documentation should be retained as required by the Medical Devices GMPS
(refer to 820.180(b)). This requirement is applicable to paper-based and electronicbased records.
2 The retention of supporting documentation such as maintenance records, change control
records, and project documentation cannot be less than 2 years from the date of release
of the medical device for commercial distribution by the manufacturer.
3 The retention of supporting documentation such as maintenance records, change control
records, and project documentation depends on the following situations:
(a) obsolete system documentationretained for the expected life of the device
supported by the legacy functionality or system
(b) retired system documentationretained for the expected life of the device
supported by the retired system
(c) current documentationretained permanently until (a) or (b) above apply
820.181 Device Master Record
Each manufacturer shall maintain device master records (DMRs). Each manufacturer
shall ensure that each DMR is prepared and approved in accordance with Sec. 820.40.
The DMR for each type of device shall include, or refer to the location of, the following
information:
(a) Device specifications including appropriate drawings, composition, formulation,
component specifications, and software specifications
(b) Production process specifications including the appropriate equipment specifications,
production methods, production procedures, and production environment
specifications
(c) Quality assurance procedures and specifications including acceptance criteria and the
quality assurance equipment to be used
(d) Packaging and labelling specifications, including methods and processes used
(e) Installation, maintenance, and servicing procedures and methods
Appendix H
265
ANALYSIS
1 There should be an individual(s) with the responsibility for preparing, dating, and
signing the master record.
2 The authorization of changes to the master records must be documented.
820.198 Complaint Files
(a) Each manufacturer shall maintain complaint flies. Each manufacturer shall establish
and maintain procedures for receiving, reviewing, and evaluating complaints by a
formally designated unit.
(b) Each manufacturer shall review and evaluate all complaints to determine whether an
investigation is necessary. When no investigation is made, the manufacturer shall
maintain a record that includes the reason no investigation was made and the name of
the individual responsible for the decision not to investigate.
(c) Any complaint involving the possible failure of a device, labelling, or packaging to
meet any of its specifications shall be reviewed, evaluated, and investigated, unless
such investigation has already been performed for a similar complaint and another
investigation is not necessary.
(d) Any complaint that represents an event that must be reported to FDA under part 803
or 804 of this chapter shall be promptly reviewed, evaluated, and investigated by a
designated individual(s) and shall be maintained in a separate portion of the complaint
flies or otherwise clearly identified. In addition to the information required by Sec.
820.198.
(e) When the manufacturers formally designated complaint unit is located at a site
separate from the manufacturing establishment, the investigated complaint(s) and the
record(s) of investigation shall be reasonably accessible to the manufacturing
establishment.
(f) If a manufacturers formally designated complaint unit is located outside of the United
States, ecords required by this section shall be reasonably accessible in the United
States.
ANALYSIS
1 There should be a designated unit to review written and oral complaints.
2 The reasons for not investigating a complaint must be documented. The authorization
of this decision must also be documented.
3 Complaints relating to injury, death or hazard to safety must be investigated
immediately.
4 Records of an investigation must be maintained.
Appendix H
266
FOOD GMP REGULATIONS RELATED TO COMPUTER

SYSTEMS VALIDATION
In the US, many regulations are applicable to food processing business. For example:
Part 210, Good Manufacturing Practice Regulations
Part 11, Electronic Records and Signatures
Part 106, Infant Formula Quality Control Procedures
Part 110, Current Good Manufacturing Practice in Manufacturing, Packing, and
Holding Human Food
Part 123, Fish and Fishery Products
Part 129, Bottled Drinking Water
The US FDA is closely evaluating other areas such as food and dietary supplements,
where proposed rulings have already been published (website address below).
A comprehensive US FDA guideline covering regulatory requirements applicable to
computer systems performing regulated function in the food industry can be found at
mailto:http://www.fda.gov/oralinspect_ref/igs/foodcomp.html.
Appendix I
Technical Design Key Practices
For all software development, both large and small, the following design and
programming techniques are consistent with current practices and may be used to
facilitate the analysis of software components in order to reduce the chances of
programming errors.
MODULAR DESIGN
A modular design is recommended, especially for moderate- to large-scale software
development efforts. Each software module should have well-defined and readily
understood logical interfaces.
Software components should be constructed using the principles of data abstraction
(requirements decomposition). If available, a high-level, object-oriented language that
supports the construction of abstract data types should be used.
The software should be hierarchically structured as a series of layers.
SOFTWARE MODULE/PROCEDURE INTERFACES

Entries to a software module or procedure should be through external calls on explicitly
defined interfaces.
Each procedure should have only one entry point and two exit points at most (one for
normal exits and one for error exits).
Data should be communicated between software modules and between procedures
through the use of procedures except where necessary for the implementation of
abstract data types. Input values should be checked for range errors using assertion
statements (if provided by the programming language used).
INTERNAL CONSTRUCTION
Each procedure should perform only a single, well-defined function.
Control flow within a single thread of execution should be defined using only
sequencing construct, with structured programming constructs being used for
conditional routines (e.g., if-then-else or case routines), and structured constructs
being used for loop routines (e.g., while-do or repeat-until routines).
If concurrent execution is employed (e.g., via multiple threads, tasks, or processes), the
software components should enforce limits on the maximum allowable degree of
Appendix I
268
concurrency and should use structured synchronization constructs to control access to

shared data.
The equivalence of variables should not be used to permit multiple memory usage for
conflicting purposes.
Robust command parsing and range-checking mechanisms should be implemented to
guard against malformed requests, out-of-range parameters, and I/O buffer overflows.
IN-LINE DOCUMENTATION
Each software module, procedure, and major programming construct should be
documented. The documentation should specify the functions performed along with a
(formal or informal) specification of preconditions and post-conditions.
Each loop should be preceded by a convincing argument (as a comment) that
termination is guaranteed.
Variable names should be used in only one context within the same procedure.
Each variable should have an associated comment identifying the purpose of the
variable and noting the range of allowable values, including if the range is
unrestricted.
If concurrency is employed, the documentation should specify how limits are enforced
on the maximum allowable degree of concurrency and how accesses to shared data are
synchronized in order to avoid (possibly undetected) run-time errors.
Index
A
Administrative procedures mapped to Part 11, 209212
Administration procedures necessary to guard data integrity, confidentiality, and availability, 209
210
physical safeguards, 211
technical security services, 212
technical security mechanisms necessary to guard against unauthorized access to data
transmitted over a communication network, 213
Acquisition process, 147
development process, 149150
maintenance process, 150151
operation process, 150
supply process, 148
Applicability of a computer validation model, 167172
sample situation, 168172
C
Change management, 8789
process, 8889
Compliance Policy Guide (CPG) 7153.17, 125127
adequacy/timelines of the corrective action plan, 127
impact on product quality/data integrity, 126
nature/extent of the device, 125126
Computer systems regulatory requirements, 219238
analysis, 219238
Computer validation management cycle, 2528
validation guidelines, 2628
validation policies, 26
Computer validation program organization, 934
computer system validation executive committee, 30
computer system validation groups and teams, 3132
CSV cross-functional team, 3031
management group, 32
organizational model, 29
validation program coordinators, 3234
Computer systems validation process, 3541
development life cycle and testing framework, 36
system development files, 4041
Criticality and complexity assessment, 173181
Index
270
deliverables matrices, 177181

validation model decision tree, 174177
E
Electronic record, 129135
audit trails applicable for electronic records, 131
electronic record authenticity, 134135
events, 132133
how should Part 11 records be managed 130131
instructions, 132
minimum record retention requirements, 131
preservation strategies, 133134
storage, 135
Electronic signatures, 137140
digital signatures, 138140
general concept, 137138
password-based signatures, 138
F
FDA, 1319
See also USA regulatory requirements for computer systems
Future, 153155
computer validation, 153155
G
Glossary of terms, 157164
H
hardware/software suppliers qualification, 107109
I
Introduction 21 CFR Part 11, complete guide to international compliance, 13
Inspection and testing, 4955
black box testing, 5254
documents inspections and technical reviews, 5051
other testing types, 5455
regulatory guidance, 4950
white box testing, 5152
M
Maintaining state of validation, 111115
business continuity, 114115
periodic review, 114
ongoing verification program, 115
operational management, 112114
security, 111112
Index
271
N
New computer systems validation model, 2124
O
Operational checks, 121124
Operation sequencing, 121122
Part 11 related operational checks, 122124
validation operational checks, 124
P
Part 11 remediation project, 117120
corrective action plan, 119
evaluation of systems, 118
remediation, 119120
Q
Qualification, 5780
configurable software qualification, 7678
custom-built system qualification, 7880
hardware installation qualification , 5861
operating system and utility software installation verification, 6970
Part 11 areas of interest, 70, 7273, 7576
related product for ISO/IEC 12119 IEEE standard adoption of ISO/IEC 12119, 7375
software installation qualification, 6164
standard instruments, micro controllers, smart instrumentation verification, 7072
system operational qualification, 6467
system performance qualification, 6768
standard software packages qualification, 73
R
Relevant procedure controls, 8586
S
Sample audit checklist for closed system, 215218
electronic signature and electronic records, 215218
Sample development activities grouped by project periods, 183208
computer system incidents, 206207
conceptualization period, 183
conduct a site acceptance test, 199204
development period, 183195
preparation of the qualification protocols, 195199
retirement, 204206
system life cycle, 208
Security, 93104
application security, 9899
network security, 9798
other key elements, 99104
physical security, 96
Index
272
security for computer system, 9496

SLC documentation, 8183
Source code, 105106
T
Training, 9192
in the regulated industry, 9192
Technical design key practices, 239240
Technologies supporting Part 11, 141146
data encryption, 142145
digital signatures, 145146
hash algorithms, 142
paper based versus electronic based solution, 141
U
USA regulatory requirements for computer systems, 1317
See FDA
food industry, 1819
medical devices software, 1718
V
Validation overview, 511
computer systems validation for low- criticality or low complexity projects, 11
introduction to computer systems validation process, 911
key project elements, 8
what is a computer system, 5
what is a computer systems validation, 56
which system should be validated, 89
why do we validate computer systems, 67
Validation project plans and schedules, 4347
mandatory signatures, 45
project schedule, 4547
validation project plans, 4345

21 CFR Part 11 Complete Guide To International Computer Validation Compliance PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

21 CFR Part 11 Complete Guide To International Computer Validation Compliance PDF

Uploaded by

Copyright:

Available Formats

21 CFR Part 11: Complete Guide to

International Computer Validation

21 CFR Part 11: Complete Guide

Boca Raton London New York Washington, D.C.

This edition published in the Taylor & Francis e-Library, 2005.

International Standard Book Number 0-8493-2243-X (Print Edition)

Chapter 2 Validation Overview

Chapter 3 USA Regulatory Requirements for Computer Systems

Chapter 4 New Computer Systems Validation Model

Chapter 5 Computer Validation Management Cycle

Chapter 6 Computer Validation Program Organization

Chapter 7 The Computer Systems Validation Process

Chapter 8 Validation Project Plans and Schedules

Chapter 9 Inspections and Testing

Chapter 11 SLC Documentation

Chapter 12 Relevant Procedural Controls

Chapter 13 Change Management

Chapter 16 Source Code

Chapter 17 Hardware/Software Suppliers Qualification

Chapter 18 Maintaining the State of Validation

Chapter 19 Part 11 Remediation Project

Chapter 20 Operational Checks

Chapter 21 Compliance Policy Guide (CPG) 7153.17

Chapter 22 Electronic Records

Chapter 23 Electronic Signatures

Chapter 24 Technologies Supporting Part 11

Chapter 25 All Together

Chapter 26 The Future

B Abbreviations and Acronyms

C Applicability of a Computer Validation Model

D Criticality and Complexity Assessment

E Sample Development Activities Grouped by Project Periods

F Administrative Procedures Mapped to Part 11

G Sample Audit Checklist for a Closed System

H Computer Systems Regulatory Requirements

I Technical Design Key Practices

professional needs to be aware of quality engineering, project management, and computer

For Lizette, Mikhail, Istvn, and Christian

Federal Register: December 19, 2001 (Vol. 66, No. 244).

Given the type of software commonly found in FDA-regulated operations, the

WHAT IS A COMPUTER SYSTEMS VALIDATION?

R.R.Herr and M.L.Wyrick, A Globally Harmonized Glossary of Terms for Communicating

confirmation by examination and provision of objective evidence that computer system

Figure 21. A Computer System and

WHY DO WE VALIDATE COMPUTER SYSTEMS?

Figure 22. Relative Cost of

B.W.Boehm, Software Engineering, IEEE Transactions on Computers, 1976.

KEY PROJECT ELEMENTS

availability of procedural controls to maintain the operating environment of the system

WHICH SYSTEMS SHOULD BE VALIDATED?

INTRODUCTION TO THE COMPUTER SYSTEMS VALIDATION

Classical Life CycleSystem Engineering, Analysis, Design, Code, Testing and

Figure 23. System Life Cycle

R.R.Herr and M.L.Wyrick, A Globally Harmonized Glossary of Terms for Communicating

early operational life

COMPUTER SYSTEMS VALIDATION FOR LOW-CRITICALITY

USA regulatory requirements for computer systems

manufacturing processes are configurable software. Many of the custom-built

Table 31. US Drugs GMP.

USA regulatory requirements for computer systems

Table 32. Reconciliation GMPs, EU Annex 11

Generation of accurate and complete

Use of operational system checks